Who Holds What a Firm Learns
Haute Lumière · The Gain Was Always Shared · Chapter One
Chapter One
Who Holds What a Firm Learns
Haute Lumière · The Gain Was Always Shared · Chapter One
In most organisations, somewhere tonight, a person who does the work knows something that would make the work better and is not going to say it. The knowledge is small and specific. A machine that drifts out of true after about four hours. A form that loses the one line the next shift needs. A report that three departments produce every month and none of them reads. The person knows it the way a cook knows which burner runs hot: not as a theory, but as a fact of the body, learned by being there.
They are not going to say it for ordinary reasons. The last time someone said something like it, nothing happened. Or something did happen, and it happened to the person who spoke. Or there is simply no place to say it, no meeting where it belongs, no form that asks, no one whose job it is to listen. So the knowledge stays where it was made, in a coat pocket, and goes home at the end of the shift.
This book is about the oldest serious attempt to change that, and about what seventy years of practice and research have shown about when it works. The attempt is called gainsharing. In the textbooks it is a kind of group incentive: a plan under which people in a unit share in the value of improvements in their unit's performance. That description is accurate and it leaves out almost everything that matters. The first plans were not designed as incentives. They were designed as a way of asking the people who did the work what they knew, answering them, and admitting, in money and in structure, that what they knew was worth something.
The book's argument can be put in one sentence. A gainsharing plan is a circulation before it is a payment: knowledge moving towards the people who can change the system, and answers, visibility and a share of the gain moving back. Where the circulation runs, the evidence is encouraging. Where only the payment runs, the evidence is mixed, and the plans tend to die.
Ammanuel Santa Anna's Ecstatic Economics is a long argument that economic life can be organised around vitality rather than extraction, and that this is not a sentimental preference but a more accurate account of where value comes from.1 Its chapter on the workplace states the premise directly, writing of an outdoor clothing company whose practices it admires, that they are expressions of a coherent economic philosophy:
vitality is the source of value, not an obstacle to it.2
That claim is easy to say and hard to make operational. A leader can agree with it on a Sunday and run a meeting on Monday in which nobody below a certain grade speaks. What the claim needs is structure: a way of organising ordinary work so that the people closest to it are asked, are answered, and share in what their knowledge produces.
Gainsharing is one of the few such structures with a long, documented record. It began in a failing steel mill in the late 1930s. It was championed at the Massachusetts Institute of Technology in the 1950s. It was studied through the 1980s and 1990s in manufacturing plants, a package delivery firm, an electric utility, a freight forwarder and, later, hospitals. The record includes careful single-site studies with years of monthly data, a few field experiments with control groups, large surveys of users, and frank accounts of plans that failed. It is, in other words, one of the better-evidenced ideas in the whole family of participatory economics, and one of the least understood.
It belongs in this house because it answers a question Ecstatic Economics raises and leaves to its readers: what does a vitality-centred firm do on an ordinary Tuesday? The book's own design principles for such a workplace include one that gainsharing turns into practice almost word for word. Under the heading "Feedback richness", it says:
Vitality requires information.2
A firm that does not hear what its people know is a body that cannot feel its own hands. Gainsharing, at its best, is a way of restoring the feeling.
There are three reasons this book is written now rather than at some other time, and none of them is a forecast.
The first is that the work most people do has changed shape. When Joseph Scanlon proposed his plan, most of the people it covered stood at machines, and the knowledge he wanted to hear was about steel, scrap and hours. Today a larger share of work is service, care, coordination and software, where almost all the value lies in knowledge the worker holds and the manager cannot see. The case of a package delivery firm in Chapter 6, where the first thing a gainsharing plan changed was how much people talked to one another about the job, is closer to most readers' working lives than the mill. The older idea fits the newer work better than it fitted the old.
The second is that the evidence is now old enough to read whole. The main studies were published between 1984 and 2008. Enough time has passed to see which findings were repeated and which were not, which plans survived and which ended, and where the research itself was weakest. A leader in the 1990s had to take the literature on trust. A leader now can read it at a distance, with its limits in view, and Chapters 5, 6, 7 and 9 try to do exactly that.
The third reason is that the language around shared reward has drifted towards formula. Most of what a leader will find if they search for the subject is about base periods, splits, caps and payout schedules. Those questions are real and they belong to each organisation and the people it would cover. This book does not describe them, for two reasons. The first is that no formula transfers: the right design depends on the work, the people, the measure and the history of the place. The second is the book's argument itself. The research keeps finding that the features that predict whether a plan works are mostly about participation, not arithmetic. A book that spent its pages on formulas would be spending them where the evidence says they matter least.
Who this book is for
Haute Lumière · The Gain Was Always Shared · Chapter One
It is written for four readers, and each will find a different centre of gravity in it.
A leader of a unit: a plant, a clinic, a branch, a department, a team of forty. You have authority over how work is organised and some influence over how people are paid, and you suspect that your people know more than your reports tell you. The field manual in Chapter 8 is written for you, week by week, with the decisions named and who makes each one. Chapters 2 to 5 give you the history and the evidence you will need when someone asks whether this has ever worked.
An owner or board member. You are being asked to approve something that sounds like a bonus scheme and may be something else. Chapter 3 gives you the fair critiques, Chapter 7 the cases where plans went wrong and why, Chapter 9 the means of knowing whether a plan is working, and Chapter 10 the conversation you will probably have with your management about it.
A union officer, a shop steward, an employee representative, or simply someone on the floor. Gainsharing began on your side of the table. Joseph Scanlon was a union officer when he made his first proposal, and the research in Chapter 4 suggests that plans carry more and better knowledge where unions take part in running them. Chapter 10 includes the conversation from your side.
One person thinking about their own money and work. The same circulation that runs through a plant runs through a household, a freelance practice, a marriage with a shared budget. Chapter 11 takes the idea to that scale, because the habits it asks for, asking, answering, making one number visible and returning what is saved, are habits a person can practise alone or with the people they live with, without anyone's permission.
It will not tell you that gainsharing raises productivity by a given amount. The studies report figures, and this book gives them exactly, with their sources and their limits. But the figures come from firms that chose to adopt a plan, studied by people who were often sympathetic to it, published in a literature that over-represents success. None of them is a forecast for your organisation. Chapter 9 explains how to produce the only figure that is: your own, measured honestly, against a baseline you set in advance.
It will not describe how to calculate or divide a gain. Those designs belong to each organisation and the people it covers, decided together and approved by them.
It will not pretend that participation is easy. Ammanuel Santa Anna's book, which treats participatory structures with evident sympathy, names the risk of pretending in a sentence about cooperatives and steward ownership that applies here without alteration: anyone who says such structures are easy is selling something.3 Several chapters of this book are about failure, because the people who wrote down why plans failed left the most useful record in the field.
And it will not separate the structure from the people inside it. Santa Anna's account of the somatic dimension of economic life, the idea that economic arrangements live in the nervous system as well as in contracts, is the thread that runs through every chapter. It begins with a claim about the reader:
The economy you live in is not just a system of policies, institutions, and market mechanisms.4
The sentence continues by saying that it is also a pattern in the body. A gainsharing plan asks people to change a pattern of that kind: to say what they know to people who have authority over them, and to trust that the answer will not cost them. Structure can make that safer. It cannot make it automatic. The chapters that follow treat both halves seriously.
How the book is arranged
Haute Lumière · The Gain Was Always Shared · Chapter One
The book has four movements and a workbook.
The idea and its record. Chapter 2 tells the origin: a mill, a union officer, a committee system with a shared measure attached, and the myth that gainsharing began as a bonus. Chapter 3 reads the five fair critiques, from the free-rider problem to the studies that see only survivors, each with a way through. Chapter 4 reads gainsharing through the house's lens, as a circulation in a living system. Chapter 5 works through the best-documented cases figure by figure, and builds a composite plant, labelled as one, to show the first year in order.
Harder ground. Chapter 6 reads the service firm where the first return of a plan was talk, and the field experiment in which the commitment outlived the bonus. Chapter 7 reads three harder rooms: a plan that broke its promise, the six Scanlon plans studied for their power games, and hospital programmes in which most of the measured saving came from somewhere other than the floor.
Practice. Chapter 8 is the field manual: the whole sequence, from the first conversation to the first review and the plateau after it, with roles and decisions named. Chapter 9 is about measurement: what to count, how to read a change honestly, and how to become your own study. Chapter 10 is about the hard conversations, with boards, unions, supervisors, partners, families and sceptics, each with a script. Chapter 11 takes the idea home.
The whole. Chapter 12 places gainsharing beside its neighbours, employee ownership, reciprocity and the commons, says plainly what it cannot do alone, and gives the living-systems reading of the whole in one image.
The workbook. Every Tool from every chapter is gathered at the end, with a ninety-day plan that a leader, a team or a household can begin on a Monday.
Each chapter ends with a Tool written out in full: a one-page instrument for doing what the chapter describes. They are meant to be filled in, not admired. A Tool with nothing written on it has not been used.
Every study cited in this book has been checked against the scholarly record: its authors, title, year and journal confirmed through Crossref, and its findings stated no more strongly than its own abstract or text supports. Where a source could be read only by title or abstract, the note says so. Every note also says what its source cannot show, because a figure without its limit is an advertisement, and this is not one.
Every passage quoted from Ammanuel Santa Anna's Ecstatic Economics and The Luminous 100 has been checked word for word against the text. Where a passage makes an empirical claim, such as a cost figure or a ranking, the claim is the book's and is marked as such; this book does not adopt figures it has not traced.
People named in case chapters are named as their sources name them. Every illustrative person and organisation that is not real is labelled as a composite or as hypothetical where it appears.
Return to the person with the card in their coat pocket. The question this book asks is not how to pay them for the card. It is why the card is in the pocket at all: what about the way their organisation is built made silence the reasonable choice, and what would have to change for speaking to become the reasonable choice instead.
Ecstatic Economics describes the experience of a citizen in a democracy that asks nothing of them between elections as the experience of being governed at rather than with. Its phrase for it is exact:
a felt sense of being governed at rather than governing together5
Most people at work know that feeling. Gainsharing, in its first and best form, was a way of turning the preposition round. Not by making everyone a manager, and not by paying everyone a bonus, but by building a route along which what people know can travel, and a promise that something will come back along it.
The companion books in this series read the same question from other directions: who owns the firm, how people answer generosity and unfairness, how groups govern what they share, and how the horizon of capital shapes what a firm keeps. This book stays with the most practical of them: the plant, the clinic, the team, the household, and the card.
Use this before reading further, alone or with two colleagues. It takes twenty minutes. It is not a survey of your people; it is a reading of your structure. Keep the page; the ninety-day plan in the Appendix returns to it.
1 · What is known and not said
Three things the people doing this work probably know about cost, quality, waste or safety that have not reached anyone who could act on them: · ·
How you know, or why you suspect it:
2 · The routes
Every route by which an idea from the floor can reach a decision today (a person, a meeting, a form, a channel):
For each route, who stands at the gate:
Which route was used most recently, and when:
3 · The last time
The last idea from someone without authority that you know reached a decision:
What happened to the person who offered it:
4 · Why the card stays in the pocket
Tick the reasons that apply here: no route · no answer comes back · speaking has cost people before · the idea would expose a manager · the idea might remove a job · nobody can see whether ideas change anything · other:
5 · The first promise you could keep
One promise about answering ideas that you could keep from next Monday without anyone's permission:
Chapter Two
The Mill That Asked First
Haute Lumière · The Gain Was Always Shared · Chapter Two
Picture a tin-plate mill in Mansfield, Ohio, in the late 1930s. The order book is thin, the equipment is old, and the company is losing money in a decade when losing money usually ended with the gates locked. A union officer who has also worked as a cost accountant goes to the owners with a proposal that is, on its face, strange. He does not ask for higher wages. He asks for permission to put the workers' ideas about cost and waste in front of management, formally, through committees, and for management to listen.
The officer was Joseph Scanlon. The company, in the account most later writers repeat, was the Empire Steel and Tin Plate Company. The fullest modern account of his life, Daniel Wren's 2009 biography in the Journal of Management History, was built from archives, unpublished sources and correspondence with Scanlon's daughter and a former colleague, and it describes the plan as his own synthesis of experience: union–management cooperation, employee participation and the sharing of gains from cost savings.6 The particulars of the mill story come to us through his colleagues and later writers, and this chapter tells them as they are told, without claiming more precision than the sources give. And the thing that came out of that mill, a decade later, would carry his name and be described in textbooks as a group incentive plan: a way of paying people a bonus when labour costs fall against the value of what they produce.
That description is accurate about the later plan and misleading about the first one. The bonus is not where gainsharing started. It started with a question asked of the people who did the work, at a moment when the firm could not afford not to ask. The money came after, as a way of keeping the question honest.
This chapter is about that order, because the order is the idea. A leader who adopts the formula and skips the question has bought the later plan's paperwork without the thing that made it work. And the evidence gathered over the following seventy years, which the chapters that follow read in detail (Chapter 3), keeps returning to the same finding the mill found first: participation is not a garnish on the bonus. It is most of the meal.
The best primary source for Scanlon's own thinking is short. In 1948, in the second volume of the Industrial and Labor Relations Review, he published "Profit Sharing under Collective Bargaining: Three Case Studies".7 By then he had left the mill, worked for the United Steelworkers, and joined the Massachusetts Institute of Technology, where Wren's biography shows his colleagues, Douglas McGregor above all, giving his ideas the setting in which they spread.6 The 1948 article is a practitioner's report. It describes cases rather than proving anything about them, and it should be read that way: as the inventor's own account of what he thought he was doing, not as evidence that it worked.
What it shows, read alongside the essays his colleague Frederick Lesieur collected after his death, is that Scanlon treated two things as inseparable.8 The first was a structure for ideas: production committees in each department, a screening committee above them, and a promise that every suggestion would receive an answer. The second was a shared measure: a simple ratio of payroll to the value of production, fixed from a period everyone could inspect, so that when the people on the floor found a cheaper or better way to work, the saving was visible and some of it came back to them.
The second element is the one that travelled. It is easy to explain, easy to put in a contract, and easy to copy. The first is slow, local and dependent on whether a supervisor actually reads the suggestion slip. Douglas McGregor, Scanlon's colleague at MIT, championed the plan, and in The Human Side of Enterprise in 1960 he treated it as an expression of a management philosophy rather than as a pay scheme.9 That is the right reading, and it is the reading the next sixty years kept forgetting.
Ammanuel Santa Anna's Ecstatic Economics has a sentence for the gap between the two halves. It opens his chapter on practical applications, and it could have been written over the mill gate:
A theory that cannot be practiced is a poem. A practice without theory is a reflex.10
The bonus formula, taken alone, is the reflex: a payment that fires when a number moves, with no account of why the number moved or who moved it. The committees, taken alone, are the poem: a lovely structure for voice with nothing to show that voice changed anything. Scanlon's contribution was to insist on both at once.
By the 1970s, writers on pay had settled on three "classic" gainsharing plans. They are worth knowing by what each one chose to measure, because the choice of measure is a statement about what the firm believes it is sharing.
The Scanlon Plan measured labour cost against the sales value of production. Its signature was not the ratio but the committee system around it. In most accounts the bonus was paid to everyone in the unit, not to the individual who made the suggestion, which is a deliberate choice: an idea that saves money in the paint shop usually depends on someone in shipping noticing a problem, and individual rewards teach people to hoard what they notice.
The Rucker Plan, associated with the consultant Allen W. Rucker, measured labour against value added: the value of production minus the cost of materials, supplies and services bought in. That is a more demanding measure. It rewards people for saving on materials and energy as well as on hours, and it asks them to understand the economics of the whole operation rather than their own department. The surviving descriptions of Rucker's work are mostly in practitioner texts and pay handbooks rather than in a primary article of his own; a search of the scholarly record for this book found no primary article of his with a DOI, and the details given here are the ones the standard secondary descriptions agree on.
Improshare, created by the industrial engineer Mitchell Fein, measured hours. It compared the hours actually worked against the hours an engineered standard said the output should take, and shared the difference. Its name is a compression of "improved productivity through sharing". Improshare was designed to be simpler than Scanlon and Rucker. It did not require the firm to open its financial accounts, and in its original form it did not require a committee system at all.
That last point matters for what comes next. Of the three, Improshare is the one that most completely separated the shared measure from structured participation. It is also the one that was studied most carefully at scale, which is how its basic design can be confirmed from a primary study rather than from a handbook. Robert Kaufman's 1992 analysis in the ILR Review, based on 112 firms that answered a 1988 survey sent to every known user, found a median productivity gain of about eight per cent in the first year and cumulative gains of seventeen and a half per cent by the third year, after which they levelled off.11 A large part of the gain came from fewer defects and less downtime. Smaller firms gained more than larger ones.
Those figures are real and they are often quoted. What they cannot show is what they are usually used to show. The firms answered a questionnaire about their own programme; a firm whose plan had been abandoned was less likely to be a "known user" in 1988 and less likely to answer. Kaufman's sample is a portrait of the survivors. And the finding that gains came through defects and downtime, rather than through people simply working faster, points back towards the mill: the improvement is in how work is organised, which is knowledge the people doing it hold.
The myth worth correcting
Haute Lumière · The Gain Was Always Shared · Chapter Two
The common account of gainsharing runs like this. Firms wanted productivity. Economists and engineers designed group bonuses to create an incentive for it. Some worked, some did not, and the ones that worked were the ones with a well-designed formula.
The history runs the other way. The plan began in a crisis as a participation system, at a union officer's initiative, with the bonus added as the mechanism that made participation credible. It was championed at MIT by a psychologist of management, not by an economist. And the most careful case reviews, which this book reads in Chapter 5, find that the features that predict success are mostly about participation: whether employees were involved in designing the plan, whether there were formal structures for their involvement, and whether they were favourable towards it.12
The correction matters for one practical reason. A leader who believes the myth will spend the design effort on the formula: the base period, the split, the payout frequency, the cap. Those questions matter. But the myth sends attention to the part of the plan that is easiest to copy and least likely to explain whether it works.
Santa Anna's diagnosis of what goes wrong in conventional firms is blunt about where that attention usually lands. His book describes the extractive workplace's view of people in one sentence:
The phrase "human resources" tells you everything you need to know about how extractive economics views the people inside its organizations.13
A gainsharing plan designed as a formula alone is still speaking that language. It treats the people on the floor as inputs whose output can be raised by adjusting the price paid for it. The plan Scanlon described treats them as the part of the firm that knows things management does not, and the bonus as the firm's way of admitting it.
There is a reason the first plan appeared where it did. A firm that is doing well has little need to ask its employees how to save money; it can carry its inefficiencies. A firm that is about to close has run out of other options, and its managers are suddenly willing to hear what they would ordinarily screen out.
The same pattern shows up later in the evidence. Dong-One Kim's study of 211 organisations with gainsharing experience, published in 1999, found that programmes were more likely to survive when they were approved by an employee vote, when a union took part in running them, when the firm relied less on outside consultants, and when it was in sound financial health.14 Read together with the origin, that last factor is interesting. Crisis can open the door to participation; it cannot keep it open. A plan born in desperation and run by a firm that recovers has to find a reason to keep asking once the pressure is gone.
Santa Anna's book has a line about the moment before that recovery, written about the economy as a whole rather than about one mill, that fits the scene in Mansfield closely. Of the pain built into ordinary economic life, the sense that this is simply the cost of progress, he writes:
It isn't. It's the price of a wound we haven't named.15
The mill's wound was nameable. Labour and management had spent the decade treating each other as the cause of the firm's decline. Scanlon's proposal named a different cause: that the firm was not using what its own people knew. The committees were a way of using it. The ratio was a way of proving that it had been used.
The finer points, for a leader reading this today
Haute Lumière · The Gain Was Always Shared · Chapter Two
The history is not only history. A leader who runs a team, a plant, a clinic or a firm can take four specific lessons from the first plan before anyone mentions a bonus.
Start with the answer, not the scheme. The Scanlon committees worked, where they worked, because every suggestion got a reply. Schuster's longitudinal study of a Scanlon plant reported that its production committees generated more than two thousand four hundred and seventy-seven suggestions, of which seventy per cent were implemented.16 Seventy per cent is not a sign that the workforce was unusually gifted. It is a sign that someone was reading the slips and that people learned it was worth writing them. Before any shared measure exists, a leader can make that promise: every idea answered within a fixed number of days, in writing, with the reason if it is declined.
Measure what people can actually affect. Scanlon's ratio, Rucker's value added and Fein's hours each reflect a judgement about what the people in the unit could change. A measure that moves for reasons nobody inside the unit controls, such as commodity prices or a contract won by the sales team in another city, teaches people that the number is weather. Before sharing anything, ask of the measure: if the people in this room changed how they work, would this number move, and would they be able to see why?
Choose the group deliberately. The founders shared across the whole unit. Later writers argued about whether smaller groups produced stronger effects; Kim's survival study found that a small bonus group was associated with programme survival.14 The finer point is not that small is better. It is that the group should be the set of people whose cooperation produces the gain. Draw it too small and you reward one shift for savings another shift made possible. Draw it too wide and the line between effort and result disappears.
Keep the books open enough to be believed. The Scanlon and Rucker plans both required the firm to show its people a number they could check. That is uncomfortable for many owners. It is also the condition of trust. A shared measure that people cannot inspect is a promise that the firm will tell them how well they did, and people do not change how they work for a promise of that kind.
None of these four requires a pay change. All four are the first plan's first half. A firm that does them well is in a position to consider the second half; a firm that does not will find that a formula imposed on top of silence produces mostly suspicion.
The chapters that follow take this history forward. Chapter 3 reads the critiques and the failures, including a close study of six Scanlon plans that went wrong for reasons of power rather than design. Chapter 4 reads gainsharing through the house's own lens, as circulation in a living system. Chapter 5 works through the published cases with every figure traced to its source, and Chapters 6 and 7 add cases from service work, from a hospital and from a plan that broke its promise. The rest of the book is practice: the field manual, measurement, the hard conversations, and the household.
What the origin establishes is a test that every later chapter applies. When someone describes a gainsharing plan, ask what it does before it pays anything. If the answer is nothing, it is the later plan's paperwork. If the answer is that it asks, answers and opens the books, it is the idea Scanlon carried out of the mill.
Santa Anna's book frames the whole field of practical economics with the same impatience about ideas that never become structure:
But groundwork that never becomes building is just a particularly eloquent hole in the earth.10
Scanlon's committees were building. They were small, procedural and unglamorous, and they are the part of the history that deserves to be copied first. The mill in Mansfield is long gone. What survives of it is a practice that cost nothing to start: a slip of paper, a committee that read it, and an answer that came back.
Use this before any conversation about shared rewards. It takes one hour, with the leader and two people who do the work. Write the answers down; they become the baseline the later tools in this book return to.
1 · The question we are asking
What do the people doing this work know about cost, waste or quality that management does not? Write three examples, in their words:
When was the last time one of those examples reached a decision-maker? · What happened to it?
2 · The answer we owe
How is an idea submitted today? (none / informal / a form / a committee)
Median days from submission to a written reply, measured from the last ten ideas:
The promise we can keep from Monday: every idea answered in writing within ___ days, with the reason if declined. Owner:
3 · The measure people can see
One number this group could move by changing how it works:
Three things that move it which this group does not control:
Can the people in the group see this number today? yes / no · How often?
What would have to be opened (a report, a ledger, a meeting) for them to check it?
4 · The group
Who cooperates to produce a change in that number? List the roles:
Who is left out of that list but affects the number?
Is the group drawn around cooperation or around the org chart?
5 · Before any money
If fewer than four are ticked, the next step is the unticked item, not a pay scheme. Write it:
Review date, ninety days from today:
Chapter Three
Where the Gain Goes Missing
Haute Lumière · The Gain Was Always Shared · Chapter Three
In 1996 the management scholar Denis Collins published a short article in Compensation and Benefits Review with a title that does not bury its subject: "Fifteen Lessons Learned from the Death of a Gainsharing Plan".17 Two years later he published a book, Gainsharing and Power, built on close study of six Scanlon-type plans, with a chapter on what he called the power games that decide their outcomes.18
That a scholar of the field would write about a plan's death so plainly is worth noticing. Most of the literature on gainsharing is written by people who believe in it, and it is a credit to that literature that it has kept a record of its failures. Chapter 2 told the origin: a failing mill, a union officer's question, a committee system with a shared measure attached. This one reads the other side of the ledger.
The aim is not to argue against gainsharing. It is to name the five ways it most often goes wrong, fairly, and to say for each one what a leader can do about it before the first meeting. A practice that has been studied for seventy years has earned that kind of reading. Ammanuel Santa Anna's Ecstatic Economics, which treats participatory structures with obvious sympathy, sets the tone in a passage about the hard work of building them:
Anyone who tells you otherwise is selling something.19
He is writing about cooperatives and steward ownership, and about the claim that such models are easy. The sentence applies here without alteration.
The sharpest theoretical objection to any group reward is old and simple. If two hundred people share a gain equally, a person who works harder receives one two-hundredth of the extra value their effort creates. A person who coasts loses one two-hundredth of what they withheld. In a firm of any size, the arithmetic says, the incentive to make a special effort for a shared bonus is close to nothing.
Eugene Kandel and Edward Lazear set this out in the Journal of Political Economy in 1992, in an article titled "Peer Pressure and Partnerships".20 Their abstract states the problem in a sentence: in organisations of any significant size, the free-rider effects would seem to choke off any motivational forces. And the same author, Lazear, later published one of the best-known findings in favour of the opposite approach. When the Safelite Glass Corporation moved its windshield installers from hourly pay to individual piece rates, output per worker rose by about forty-four per cent.21
That is a serious challenge, and it should be taken seriously rather than waved away. But Kandel and Lazear did not stop at the problem. Their article is about how peer pressure works, and how profit sharing, shame, guilt, norms, mutual monitoring and empathy interact to create incentives that the simple arithmetic leaves out. The free-rider problem is real when people are strangers to each other's work. It weakens when they can see each other's work, care what their colleagues think, and share a norm about what a fair day looks like.
The Safelite result also needs reading carefully. Installing a windshield is a task one person does alone, with output that can be counted exactly. Individual piece rates suit that kind of work. Most of the gains described in the gainsharing literature come from a different kind of work: fewer defects, less downtime, ideas about how a process could run better. Those are rarely the product of one person's effort, and a piece rate cannot pay for them.
The way through. Draw the group around the people who can see each other's work, and keep it as small as the cooperation allows. Dong-One Kim's survey of 211 organisations found that a small bonus group was among the factors associated with a plan's survival.22 Then build the visibility on purpose: a shared board, a weekly look at the measure, a habit of naming who found what. The free-rider problem is a description of a group that cannot see itself. Let it see itself.
Every gainsharing plan compares present performance against a base: a period in the past, a standard, a ratio. The base is where trust is won or lost. If the base is set at a period of unusual weakness, the plan pays out for recovery that would have happened anyway. If it is set at a period of unusual strength, the plan never pays out at all. And if management resets the base upward every time the group beats it, which is the ratchet, people learn that a good year is punished with a harder target.
The ratchet is not a theoretical fear. It is the oldest complaint about piece rates, older than gainsharing, and it explains why workers under some individual incentive schemes have kept output just below the level that would trigger a new standard. A gainsharing plan that ratchets has rebuilt the problem it was meant to solve.
Theresa Welbourne and Luis Gomez-Mejia's 1995 critical review of the field in the Journal of Management treated the design of the base and the conditions mediating success as open questions, and called for research rather than claiming the matter settled.23 That is the honest position. There is no base that is right for every firm.
The way through. Agree the rules for changing the base before the plan starts, in writing, and put them to the people the plan covers. Kim found that employee approval by vote was among the factors associated with survival.22 A vote is not a guarantee of fairness, but it changes who owns the rule. A base that people helped set, and can read, is a base they will defend when the market turns. Where capital investment changes the work, say in advance how the base will be adjusted for it; Kim's study also found major capital investment associated with survival, which suggests that firms which invest alongside the plan, rather than in place of it, keep it alive longer.
The third critique: power, and the supervisor in the middle
Haute Lumière · The Gain Was Always Shared · Chapter Three
Collins's six plans are the most useful single source on the part of gainsharing that the formula cannot see. His book's chapter titles say where he looked: participatory management in Scanlon-type plans, and the power games, outcomes and lessons learned. The finding most often drawn from his work is that plans fail less often because the arithmetic is wrong than because the participation is never real. Managers keep the right to decline suggestions without explanation. Committees become places where management reports and employees listen. And supervisors, whose authority rested on knowing more than the people they supervised, find that the committee structure makes that authority visible as a bottleneck.
The supervisor deserves particular care. In a traditional plant the supervisor is the channel through which problems are reported and decisions are made. A gainsharing plan with a working suggestion system creates a second channel. Ideas arrive at the screening committee without passing through the supervisor, and some of them describe problems the supervisor knew about and did not fix. A supervisor who experiences the plan as an audit will resist it, often quietly, and the plan will slowly lose the people in the best position to make it work.
Santa Anna's book names this kind of resistance from the inside of the nervous system rather than the organisation chart. Writing about workplaces that try to move from hierarchy to distributed authority, he is precise about why structure alone does not hold:
You cannot distribute authority to nervous systems that are still organized around fear.24
A supervisor who fears the committee is acting on that fear rationally. The plan has threatened something real.
The way through. Redesign the supervisor's role before the plan starts, not after. Make the supervisor the person who brings the team's ideas to committee and is credited when they are adopted, rather than the person whose omissions are exposed by them. Measure supervisors on the share of their team's ideas answered and implemented. And keep the rule that every idea receives a written answer, with a reason, within a fixed time. That rule is the plainest test of whether power has actually moved. If the answers stop coming, the plan is already dying, whatever the bonus says.
A shared measure is affected by things the group controls and things it does not. A rise in the price of steel, a contract lost by the sales team, a recession, a new machine bought by head office: each can move a value-added measure or a labour-cost ratio by more than a year of good ideas. When that happens, the plan pays out for luck or withholds a payout despite real improvement. Either way, the people covered learn that the number is weather.
This critique is the reason Mitchell Fein built Improshare around engineered hours rather than financial ratios. Hours are closer to what the group controls. They are also narrower: an hours measure does not reward saving materials, energy or scrap, which is often where the larger gains lie. Each of the three classic designs, described in Chapter 2, made a different trade between controllability and scope. None escaped the trade.
Robert Kaufman's study of Improshare users found that productivity gains rose to a cumulative seventeen and a half per cent by the third year and then levelled off.25 A levelling-off is not a failure, but it raises a design question that the formula alone cannot answer. When the easy gains have been found, what keeps the plan alive? If the measure only rewards improvements that are now exhausted, the plan becomes a fixed cost with no visible return, and it will be cut in the first bad quarter.
The way through. Choose a measure the group can affect, and say plainly in advance which outside events will be adjusted for and how. Then plan for the plateau. Kim's survey associated retraining and new-employee training with survival, and found that these worked through their effect on programme performance.22 The practical reading is that a plan stays alive by widening what people can see and change: training that teaches the economics of the next process up the line, so that the next round of ideas has somewhere to go.
The fifth critique: the evidence sees the survivors
Haute Lumière · The Gain Was Always Shared · Chapter Three
The last critique is about the research itself, and it applies to every favourable figure in this book. Most of what is known about gainsharing comes from three kinds of source. Case studies, written by consultants or researchers who helped install the plan. Surveys of current users, which cannot easily reach firms that abandoned their plans. And a small number of quasi-experimental studies at single sites.
Each has a known weakness. Case studies are published more often when the plan worked. Surveys of known users are portraits of survivors; Kaufman's 112 Improshare firms answered a 1988 questionnaire sent to every known user, and a firm that had dropped the plan was less likely to be known or to answer. Single-site studies can be careful and still say little about the next site. Robert Bullock and Edward Lawler's 1984 review in Human Resource Management found that most of the cases in the literature were successful, and gave its article a title that is still the fairest summary of the field: "Gainsharing: A few questions, and fewer answers".26
None of this means the favourable findings are wrong. It means they are likely to overstate the average result, and that a leader should not read any figure in this book as a forecast for their own firm.
The way through. Treat your own plan as the study. Write down, before the plan starts, what you expect to change and by when, and how you will know. Collect the measure for a year before anything is shared, if you can, so that the base is a reading rather than a guess. And keep the record of what failed as carefully as the record of what worked. Collins's fifteen lessons came from a plan's death. A firm that keeps that kind of record about itself will learn more from its own plan than from any published case.
Read together, the five critiques point at the same place. Free riding is a problem of visibility. The ratchet is a problem of who owns the rules. Power is a problem of whether voice is answered. Weather is a problem of what the measure can see. Survivorship is a problem of what the research can see. In each case the formula is downstream of something else: whether the people covered can see the work, shape the rules, be answered, and trust the number.
That is where Chapter 2 left the history, and it is where Chapter 4 picks up the house's own reading. Santa Anna's book states the general principle in a line about organisations that try to change their culture without changing their structure:
Culture doesn't float free of structure.27
The critiques of gainsharing are the same sentence read from the other side. Structure does not float free of culture either. A formula installed in a firm where voice is unanswered becomes another reason for suspicion. A formula installed in a firm that has already learned to answer becomes a way of saying, in money, what the answers already said.
Two companion books in the Ecstatic Economics series reach the same conclusion from different directions. The book on employee ownership reads the evidence on what happens when people own part of the firm rather than share in a measure. The book on Elinor Ostrom's work on the commons describes how groups that share a resource make and enforce their own rules, which is the free-rider problem solved without a central authority. Both are worth reading beside this one.
Collins's article is short. Its title is the most useful thing about it: a reminder that the plans worth learning from include the ones that ended, and that someone stood by the grave long enough to write down why. Chapter 7 returns to a plan of that kind, read from the inside.
Run this before any shared-reward plan is announced. Imagine it is two years from now and the plan has been abandoned. For each critique, write the most likely reason it failed, then the step that would prevent it. One hour, with the leader, a supervisor, and two people the plan would cover.
1 · One share among many (free riding)
Where the group cannot see itself, the visibility we will build (a board, a weekly review, a naming habit):
Smallest group that still contains the cooperation that produces the gain:
2 · The ratchet and the base
Proposed base period or standard: · Was it a normal period? yes / no · evidence:
Written rule for when and how the base changes:
Will the people covered vote on the plan and the rule? yes / no · if no, why not:
How capital investment will be treated:
3 · Power and the supervisor
Each supervisor's role in the plan, in one sentence:
How supervisors will be credited when their team's ideas are adopted:
Written-answer rule: every idea answered within ___ days, with a reason if declined · Who checks it is kept:
The first sign that power has not moved, and who will report it:
4 · The number moves for other reasons
Things inside the group's control that move the measure:
Things outside it that move the measure:
How each outside event will be adjusted for, decided now:
When the easy gains are exhausted, what training will widen what people can change?
5 · The evidence sees the survivors
What we expect to change, by how much, by what date (written before the plan starts):
Where we will record what fails, and who keeps it:
The verdict
Critiques with no preventing step written:
If any, the plan waits until each has one. Review date:
Chapter Four
A Firm Has a Circulation
Haute Lumière · The Gain Was Always Shared · Chapter Four
In the early 1990s three researchers ran an unusual field experiment. Susan Hanlon, David Meyer and Robert Taylor studied a gainsharing plan with a control group, measured how employees thought and behaved while it ran, and then kept measuring after the group bonus was taken away.28 The question was what would be left.
The intuitive answer, if gainsharing is an incentive, is nothing. Remove the payment and you remove the reason for the behaviour it paid for. What they found, three months after the bonus ended, was something else: the employees who had been under the plan showed higher moral commitment, more prosocial behaviour and less intention to leave than the control group.
One study, at one site, three months out. It cannot say how long the effect lasts or whether it would appear elsewhere. But it is hard to explain if a gainsharing plan is only a price paid for effort. It is easier to explain if the plan changed something about the relationship between the people and the firm, and the change outlived the money.
That is the reading this chapter offers. Chapter 2 traced gainsharing to its origin in a participation system; Chapter 3 read the five ways it fails. This one reads it through the lens Ammanuel Santa Anna develops in Ecstatic Economics and in the Luminous 100: a firm as a living system, and gainsharing as one of the ways that system keeps its circulation open.
Santa Anna's book draws a distinction early that does most of the work in what follows. Writing about the anthropology of gift economies, from Marcel Mauss onward, he describes societies in which economic life was organised around circulation rather than accumulation, and he puts the difference in two short sentences:
Wealth was not something you had. It was something you moved.29
He is careful about the history. The same passage goes on to say that gift economies could be coercive, that the potlatch could impoverish, and that the point is to learn the principle without romanticising its setting. The principle, stated a few paragraphs later, is the one that matters for a firm:
wealth that does not circulate is not wealth. Prosperity that does not connect is not prosperity.29
Apply that to a factory, a clinic, a software team. There is a kind of wealth in every working organisation that does not appear on its balance sheet: what the people doing the work know about how it could be done better. Where the machine sticks. Which supplier's parts fail. Which handover loses information. Which step exists only because someone once needed it. That knowledge is held in bodies and habits on the floor. In most firms it stays there.
The conventional firm treats this as a motivation problem: people do not volunteer what they know because they lack an incentive. The living-systems reading treats it as a circulation problem. The knowledge does not move because there is no channel for it to move along, and because, when it has moved before, nothing came back. A blood vessel that carries nothing in one direction soon carries nothing in the other.
Read this way, a working gainsharing plan has two currents, and the bonus is only part of one of them.
The first current runs up. Knowledge moves from the people who do the work to the people who can change the system: through the suggestion, the committee, the conversation at the board. Scanlon's committee structure was a vascular system for that knowledge. It gave ideas a route that did not depend on a single supervisor's attention.
The second current runs back down. Three things return. An answer to every idea, with a reason if it is declined. Visibility of the shared measure, so people can see whether what they did changed anything. And, when the measure moves, a share of the gain. Of the three, the answer is the most frequent and the least expensive, and it is the one most firms neglect.
When both currents run, the gain is not the purpose of the plan. It is a sign that the circulation is working, in the way a healthy pulse is a sign that blood is moving rather than the reason the heart beats. And when the money stops, as it did in the Hanlon experiment, the currents do not necessarily stop with it. The channel for ideas and the habit of answering can survive the end of the bonus, because they were never only about the bonus.
Santa Anna's book has a line for this, written about self-managing organisations rather than gainsharing plans, that names the second current exactly:
trust circulates, and where trust circulates, value flows30
The order in that sentence is worth noticing. Trust first, value after. A plan that tries to make value flow by paying for it, without first building the channel that trust travels along, has the order backwards.
What the research says about the up-current
Haute Lumière · The Gain Was Always Shared · Chapter Four
The organisational learning research on gainsharing is the strongest support for this reading, because it studied the up-current directly.
Jeffrey Arthur and Lynda Aiman-Smith analysed the suggestions employees submitted over the first four years of a gainsharing plan at one plant, and proposed a model of gainsharing as an organisational learning system.31 Their interest was not only whether people made suggestions but what kind: whether, as a plan matured, ideas moved from simple fixes within an existing process towards ideas that questioned the process itself. That second kind, which learning theorists call second-order learning, is the knowledge most valuable to a firm and most risky for the employee who offers it, because it often implies that something a manager designed was wrong.
Arthur and Dong-One Kim followed this with a comparison of two organisations with similarly structured plans but different degrees of cooperation between labour and management.32 Their hypothesis was that employees' willingness to share high-risk knowledge depends on trust: that union support and involvement would increase people's initial acceptance of the risk built into a variable reward, and their willingness to take on further risk over time. The results, they report, generally supported the hypothesis.
Two plants are not a population, and a finding that "generally supports" a hypothesis is not a law. But the direction is clear and it matches the living-systems reading. The up-current carries more, and more valuable, knowledge when the channel is trusted. The same plan design produced different flows in two places because the relationships through which it ran were different.
If the plan is a circulation, its failures are blockages, and the five critiques of Chapter 3 can be read again in that light.
The unanswered suggestion is the simplest. An idea that goes up and receives nothing back teaches the person who sent it that the channel is decorative. Ten unanswered ideas teach a department. The Luminous 100, Santa Anna's upgrade of the hundred frameworks consultancies most often teach, reads Lean thinking's seven wastes in exactly these terms. Of waiting, the waste of work standing still between steps, it says:
waiting reveals a flow blockage where developmental energy is pooling33
An idea that waits three months for an answer is energy pooling. The person who had it does not stop thinking; they stop sending.
The ratchet is a blockage in the down-current. A group that improves and finds its target raised has received, as its return, a harder target. The circulation now runs one way: effort goes up, nothing comes back. The free-rider problem is a blockage of visibility: people who cannot see each other's contribution cannot respond to it. The supervisor who resists the committee is a valve closing at the point where the old structure feels threatened. And the measure that moves with the weather is a signal that no longer tells the organism anything about its own health.
This is the practical value of the reading. It turns a list of design faults into a single question that can be asked of any plan at any time: where is the current not running?
The soil and the fruit
Haute Lumière · The Gain Was Always Shared · Chapter Four
The same upgrade in the Luminous 100 offers a second image that fits gainsharing closely. Its reading of Google's Project Aristotle, the study of what made teams effective, treats the five factors the research named as parts of one living team rather than separate dials to turn, and ends on a sentence that could be posted above any incentive design meeting:
you cannot produce fruit by optimizing the fruit; you must nourish the soil.34
The gain in a gainsharing plan is fruit. It is the visible result of a team that has found a better way to work. A plan designed around the fruit alone, with care spent on the split, the cap and the payout frequency, is trying to produce fruit by attending to fruit. The soil is the circulation: the answered idea, the visible measure, the trusted channel, the supervisor who brings ideas forward rather than guarding against them.
Santa Anna's book makes the corresponding claim about structure and culture directly. Writing about benefit corporations, he argues that organisations cannot reliably produce one kind of culture inside a structure built for another, and states the conclusion without qualification:
you cannot reliably produce ventral vagal organizational culture within sympathetic-activation corporate structures.35
The language is the book's own, drawn from polyvagal theory: the difference between a nervous system in social engagement and one in defensive activation. Read for gainsharing, the claim is that a plan which asks people to share risky knowledge, inside a firm whose structure still punishes the bearer of bad news, will not get the knowledge. People in a defensive state do not volunteer that the process their manager designed is wrong. The plan has to change the structure enough that offering the idea feels safe, or the up-current will carry only the safe ideas.
The living-systems reading is only worth having if it changes what a leader does on an ordinary Tuesday. It changes five things.
It moves the first investment from the formula to the channel. Before any discussion of shares, build the route by which ideas travel and the promise that they will be answered. Measure the channel: ideas received per month, median days to answer, share implemented. These are the vital signs. A plan with no reading on them is a plan whose health is unknown.
It makes the answer the most important payment. The bonus is paid occasionally. The answer is paid every time someone speaks. A firm that answers every idea within a week, with a reason, is paying its people in the currency the Hanlon experiment suggests outlasts money: the experience of being heard by the system they work in.
It reads failure as a location, not a verdict. When a plan stalls, the question is not whether gainsharing works. It is where the circulation has blocked: the up-current (no ideas), the answer (ideas but no replies), the measure (replies but no visible change), or the share (visible change but nothing returned). Each location has a different remedy.
It treats the end of the bonus as survivable. Plans end, because firms are sold, markets turn and formulas age. A plan built on circulation can lose its bonus and keep its channel. A plan built on the bonus alone loses everything at once. If the Hanlon result holds even partly, the work of building the channel is not lost when the money stops.
It applies at the scale of a kitchen table. The same reading holds for a practice of four, a family business or a household that keeps a shared budget: the people closest to a cost usually know the most about it. Chapter 11 takes the reading to that scale. The scale changes the furniture, not the circulation.
There is one finer point that applies at every scale, and it is the one most often missed. The down-current has to reach the person, not only the group. A plant that posts "defects down eleven per cent" has returned a measure. A plant that posts "defects on line three down after the night-shift card about the sealing jaw", with the name of whoever wrote it, has returned an answer, a measure and a recognition at once, and has told everyone else on the floor that cards are read. The circulation is felt by individuals even when the gain is shared by the group.
Santa Anna's book, discussing what makes worker ownership different, describes a change in orientation that gainsharing at its best begins to produce without changing ownership at all:
This felt ownership changes everything.36
The companion book on employee ownership reads what happens when that felt ownership is made legal. The companion book on gift and reciprocity reads the experimental evidence that people answer generosity with effort, which is the down-current studied in a laboratory. Chapter 5 works through the published cases, figure by figure.
Three months after the bonus ended, in a plant whose name the study does not give, people who had been asked and answered were still a little more committed to one another than the people who had not. The money had gone. The channel, for a while at least, was still open.
Draw this on one page for any team or unit, with or without a gainsharing plan. Fill it from records where they exist and from honest estimates where they do not, and mark which is which. Repeat it every quarter.
1 · The up-current (knowledge moving towards the people who can change the system)
Routes by which an idea can travel today:
2 · The answer (the first return)
3 · The measure (the second return)
The number this team can see that tells it whether its ideas changed anything:
How often the team sees it: · Who shows it to them:
4 · The share (the third return)
What, if anything, returns to the team when the measure improves (recognition, time, training, money):
Has the target ever been raised because the team beat it? yes / no · when:
5 · Where the current blocks
The one change that would reopen that stage this quarter: · Owner:
The vital sign that will show whether it worked, and its value today:
6 · The soil
The most recent example of someone saying so, and what happened next:
Chapter Five
Twenty-One Defects in a Thousand
Haute Lumière · The Gain Was Always Shared · Chapter Five
Here is a number worth holding in the hand. At a manufacturing company studied by Larry Hatcher and Timothy Ross, defective products ran at twenty point nine three per thousand shipped while the plant paid its people by individual piecework. After the company ended piecework and introduced a plant-wide gainsharing plan with a structured system for employee ideas, the figure was two point three one per thousand.37
That is a fall of roughly nine-tenths. It was measured by interrupted time-series analysis over more than four years of objective data, which means the authors looked for a change in the trend at the moment of the intervention rather than simply comparing two averages. They also found a significant fall in grievances at the same moment, and surveys taken two months before the change and fifteen months after showed better perceptions of teamwork and more concern for performance.
It is also one company. The authors interpret it through Morton Deutsch's theory of cooperation and competition: individual piecework set workers against each other, and a shared measure set them alongside each other. That is a plausible explanation and it is theirs. The study shows that quality changed when the pay system changed; it cannot show that the same change would produce the same result in a different plant, with different products and different people.
This chapter works through the best-documented public cases in that spirit. The chapters before it told the origin, the critiques and the house's living-systems reading. This one does the arithmetic, traces every figure to its source, says what kind of evidence each is, and then puts the lessons in order in a composite plant, labelled as one, so that a leader can see the first year in sequence.
Michael Schuster's study of a Scanlon Plan in a large manufacturing plant, published in 1984, is one of the longest single-site records in the literature.38 He collected monthly data on productivity (output per hour), employment and voluntary turnover over a period of seven and a half to nine years, depending on the measure, and analysed it as an interrupted time series.
His abstract reports four findings, and each is worth reading at its exact strength.
An abrupt positive change in productivity, followed by positive trends for two product lines. "Abrupt" is a technical word here: the level of output per hour stepped up at the point the plan began, rather than drifting upward over years. A step at the moment of intervention is harder to attribute to background change than a slow drift.
Stable employment and turnover, in contrast to wider shifts in the industry. This is easily overlooked and arguably the most important finding. A plan that raises productivity by cutting jobs has moved the gain from one group to another. A plan that raises productivity while employment holds steady has created something.
Consistent bonuses. The plan paid. The abstract does not give the amounts, and this chapter does not supply them.
More than two thousand four hundred and seventy-seven suggestions from the production committees, of which seventy per cent were implemented. This is the up-current of Chapter 4, counted.
What the study cannot show is equally specific. It is one plant. Schuster's own measurement cannot separate the effect of the bonus from the effect of the committees, since both began together. And it does not say what the plant was like before the plan in terms of trust between labour and management, which the later research on knowledge sharing suggests matters a great deal.
Robert Kaufman's 1992 study of Improshare is the largest quantitative study in this set.39 He analysed 112 responses to a 1988 questionnaire sent to all known users of the plan. The figures:
Each figure has a source and a limit. The productivity figures are as reported by the firms, not measured independently. The sample is firms still known to be using Improshare in 1988 who chose to reply; firms that tried the plan and dropped it are unlikely to be well represented, and they are exactly the firms whose results a cautious reader most wants to see. The finding that the median was lower than the mean is a useful warning: if you are going to use one figure from this study to think about your own firm, use the median, and treat even that as an upper estimate.
The finding about defects and downtime is the one that best survives these limits, because it describes the mechanism rather than the size. It agrees with the Hatcher and Ross plant. Across very different designs, the gains in these studies come disproportionately from quality and from the machine standing idle less often. That is knowledge about how work is organised, and it lives with the people doing the work.
Ammanuel Santa Anna's Ecstatic Economics frames the general point in a sentence written about workplaces that change what drives performance rather than how hard it is pushed:
The vitality-centered workplace doesn't reject productivity. It redefines what drives it.40
The Improshare figures, read for their mechanism, are a small instance of that redefinition. The gain is not mostly people working faster. It is fewer things going wrong.
Thirty-three cases and a disagreement
Haute Lumière · The Gain Was Always Shared · Chapter Five
Two reviews summarise what predicts success across many plans, and they agree on most points and disagree on one, which is worth knowing.
Robert Bullock and Mark Tubbs, in 1990, gathered thirty-three published case studies of gainsharing plans implemented over fifty years and tested which features predicted success.41 The ones that did were employee involvement in designing the plan, the use of outside practitioners, formal involvement structures such as committees, and employees' favourability towards the plan. Situational conditions, with the exception of a participative management style, did not predict success. And success was robust across organisation size, union status, technology and environment.
Dong-One Kim, in 1999, surveyed 211 organisations with gainsharing experience and asked a different question: which plans survived?42 His associations included employee approval by vote, union participation in running the plan, retraining and new-employee training, a small bonus group, labour intensity, sound financial health and major capital investment. And he found survival associated with low reliance on outside consultants.
So one review finds outside practitioners associated with success, and the other finds reliance on outside consultants associated with failure to survive. Both can be true. A practitioner who helps a firm design its plan and build its committees, then steps back, may raise the odds of a good start. A firm that continues to rely on outsiders to run the plan may never come to own it, and plans that are not owned end. The finer point for a leader is in the verb. Use outside help to build the participation system. Do not use it to operate one.
Both reviews share a weakness Chapter 3 named: published cases and surviving programmes over-represent success. Both also agree on the feature that matters most, which is that the people covered by the plan take part in designing and approving it.
What follows is a composite, built for teaching. It is not a real company, and its figures were set by the writer to illustrate the sequence the research suggests, not drawn from any filing or study. Its purpose is to show the order of work in the first year, which is where most plans are won or lost. It describes the participation system only; it does not describe how any gain would be calculated or divided.
The plant. One hundred and sixty people on three shifts, packing dried goods for supermarket own-label brands. A plant manager, six supervisors, a small maintenance team, one quality technician. Margins thin, defects high enough that two customers have complained in writing.
Months one to three: the answer before anything else. The plant manager introduces one rule and nothing else. Any idea, from anyone, about cost, quality, safety or waste is written on a card or entered on a tablet by the canteen, and receives a written answer within ten working days, with a reason if declined. Supervisors are told they will be measured on one thing in this period: the share of their team's ideas answered on time. In the first month the plant receives eleven ideas. By the third it receives forty. The plant manager answers every one, some of them with "no, because", and posts the answers on the board.
Months three to six: the measure made visible. The quality technician and two volunteers from the floor choose a measure the plant can affect: defects per thousand cases shipped, broken down by line. They post it weekly, by the board where the answers go. They also start a second column: downtime minutes per shift. No money is attached. People begin to connect their cards to the numbers. A card from a night-shift operator about a sealing jaw that drifts out of alignment after about four hours is implemented; the defect line for that machine falls, and the card's author is named beside the figure.
Months six to nine: the committees. With the help of an outside facilitator for three months, the plant forms one production committee per shift and a screening committee above them, with elected members from the floor and one supervisor each. The facilitator trains the committees to read the measure and to evaluate ideas, then steps back. The answer rule passes to the committees, with the plant manager answering any idea the committees cannot decide.
Months nine to twelve: the design and the vote. Only now does the plant design a shared reward, with the committees, and put it to a vote of everyone it would cover, together with the written rules for changing the base. The design is the plant's own business and is not described here. The vote passes. The plant manager notes, privately, that the twelve months of answered cards are what made a yes possible.
What the composite is for. It puts in order four things the research finds associated with success or survival: formal involvement structures, employee involvement in design, approval by vote, and outside help used to build rather than to run. And it puts first the thing Chapter 4 argued matters most: the answer. In a real plant some of these steps would take longer, some would fail and need repeating, and a leader might reasonably compress the sequence. The order is the lesson.
The finer points of working it
Haute Lumière · The Gain Was Always Shared · Chapter Five
The cases, read together, give a leader six specific points of practice.
Collect the measure before you share it. Schuster and Hatcher and Ross could show a change at the moment of intervention because they had data from before. A firm that begins sharing on day one has no before. Twelve months is better than six; six is better than none.
Count the ideas, the answers and the implementations. Schuster's seventy per cent implementation rate is a figure any firm can produce about itself. It is the plainest available reading of whether the up-current is real. Post it where the people who wrote the ideas can see it. Santa Anna's book, writing about pay transparency, makes a claim about visibility that applies equally to a ledger of ideas:
The nervous system relaxes when it can see the landscape clearly.43
A team that can see how many of its ideas were answered and used has stopped guessing whether the system is listening.
Watch quality and downtime first. Both Kaufman's survey and the Hatcher and Ross plant locate much of the gain there. If a plan is working, these are likely to move before output per hour does.
Hold employment steady and say so. Schuster's plant kept employment and turnover stable. People will not volunteer an idea that removes their own job. If a firm cannot promise that an idea will not cost the person who has it their place, the up-current will carry only the ideas that cost someone else theirs.
Use outside help to build, then step back. The disagreement between Bullock and Tubbs and Kim resolves into this practice.
Let the people covered vote. Both reviews, in different words, put participation in design and approval near the top.
Santa Anna's book has a line about the cooperative firm's response to crisis that reads as a summary of the composite's spirit. Describing how Mondragon handled the collapse of one of its cooperatives, it says:
The crisis was metabolized by the community rather than externalized onto individuals.44
A gainsharing plan, at its best, does something similar with improvement. The gain is found by the community and returned to it, rather than taken from it.
The same sequence holds at the scale of a practice of five or a household with a shared budget, and Chapter 11 works it through at that scale. Ask before you share. Answer every idea. Make one number visible. Let the people affected help design the rule.
On the board by the canteen, in the composite plant, the card about the sealing jaw is still pinned beside the line it changed, the operator's name in pencil, the number under it smaller than it was.
Keep this as one page per month for the first twelve months of any participation system. Every entry is a count or a date, not an opinion. It is the record that lets a firm become its own case study.
Header
Unit: · People covered: ___ · Month ___ of 12 · Prepared by:
1 · The up-current
2 · The answer
3 · Implementation
4 · The measure (collected before anything is shared)
Measure chosen: · this month: ___ · same month last year: ___
Outside events this month that moved the measure, and by roughly how much:
5 · Employment
Any role removed because of an implemented idea? yes / no · if yes, where the person went:
6 · Structure
Outside help in use: none / building / operating · planned end date for outside help:
7 · One line
The most useful idea this month, in the words of the person who had it:
Chapter Six
Talk Is the First Return
Haute Lumière · The Gain Was Always Shared · Chapter Six
Most of the gainsharing literature was written about factories. The mill in Chapter 2, the Scanlon plant in Chapter 5, the hundred and twelve Improshare firms: these are places where output can be counted in cases, tonnes or defects per thousand, and where an idea about a sealing jaw can be traced to a line on a chart. A reader who runs a clinic, a law office, a software team, a call centre or a logistics desk is entitled to ask whether any of it applies to work whose product is a service, a decision or a conversation.
There is a small body of careful research that answers the question, and it comes from service firms. This chapter reads three studies from it. Two of them come from one research programme, conducted by Susan Hanlon and Robert Taylor, later joined by David Meyer, in a firm in the priority package delivery industry. The third, by Theresa Welbourne and Claudia Ferrante, comes from a firm offering customs brokerage and freight-forwarding services. None of the three is large. All three are useful, because they looked at something the factory studies mostly did not: what happens between people while a plan runs.
The answer they give is the title of this chapter. In service work, before any money moves, the first thing a gainsharing plan changes is how people talk to one another. Talk is the first return. A leader who understands that will design the first year differently, measure it differently, and be less surprised by what survives.
In 1991 Hanlon and Taylor published a study in Group & Organization Studies of a self-designed gainsharing plan adopted by a firm they describe as a Fortune Service 500 company in the priority package delivery industry.45 Two features of that description matter before anything is said about the results.
The first is the phrase self-designed. The plan was not bought from a consultant as a finished product. The firm built it. Chapter 5 read the disagreement between two reviews about outside help, and resolved it into a practice: use outside help to build a participation system, not to operate one. A self-designed plan sits at one end of that range, and the study cannot tell us whether its results depended on that fact. It is still worth noticing that the best-designed service study in this literature was of a plan the firm owned from the start.
The second feature is the design of the research. The authors report that theirs was the first gainsharing study to use control groups together with measurements taken before and after the plan began, in a quasi-experimental field design. That is a modest-sounding claim and an important one. Most of the evidence in this book's earlier chapters is either a single site measured over time or a survey of firms that still had their plans. A control group lets a researcher ask whether a change in the plan's group also happened, for other reasons, in a comparable group that had no plan. It does not make a field study into a laboratory, and the abstract does not tell us how the groups were chosen or how alike they were. But it is the right instinct, and it is rarer in this field than it should be.
What they measured was communication. Specifically, the study focused on changes in work-group communication behaviours after the plan was installed. Their finding, as the abstract states it, is that during the early stages, the first six months of what they call a successful gainsharing programme, positive job-related communication and the communication of ideas increased among the people covered by the plan.
That is the whole of the result available from the abstract, and it is enough to carry an argument. In the first six months, before a plan has had much time to pay anything, people under it began to talk more about the job and to pass more ideas to one another. The authors drew a conclusion for researchers that is equally a conclusion for leaders: that the fruitful approach is to study the intervening effects of gainsharing, the process through which it operates, rather than looking only at the economic outcomes at the end.
The house's reading, set out in Chapter 4, describes a working plan as two currents: knowledge moving up towards the people who can change the system, and answers, visibility and a share of the gain moving back. Hanlon and Taylor's result shows the up-current starting to run before the down-current has carried much money. People spoke first. The money, where it came, came later.
There is a plain reason for that order. A shared measure changes what is worth saying. Under individual pay, or under a salary unconnected to any group result, telling a colleague how you save ten minutes on a route is a favour; it may even be a small loss, if the two of you are compared. Under a shared measure, the same remark has become a contribution to something you both hold. The plan has not made anyone more generous. It has changed the meaning of a sentence.
Ammanuel Santa Anna's Ecstatic Economics describes the condition that makes this possible in its list of design principles for what it calls the vitality-centred workplace. The fourth principle is feedback, and the book states the cost of its absence in the language of a body:
Organizations that suppress feedback — through hierarchy, blame cultures, or information hoarding — create the economic equivalent of a body that can't feel pain46
A delivery firm whose drivers, sorters and dispatchers do not tell one another what they notice cannot feel where its routes hurt. The finding from Hanlon and Taylor is, read this way, the first evidence of sensation returning. More talk about the job is more information moving through the organism. More ideas exchanged is the organism beginning to correct itself.
The finding also explains something the factory studies could only infer. Chapter 5 noted that much of the measured gain in the Improshare firms and in the Hatcher and Ross plant came from fewer defects and less downtime rather than from people working faster. In a factory, it is easy to assume those gains came from individual insights: one operator, one machine, one card. The service study suggests a broader mechanism. The gains may come from people talking to each other about the work more often, so that what one of them knows reaches the person who can use it. In service work, where the product is coordination, that is very close to being the product itself.
When the bonus was taken away
Haute Lumière · The Gain Was Always Shared · Chapter Six
The second study in the programme is the one Chapter 4 opened with. Hanlon, Meyer and Taylor followed the same kind of plan through an event most researchers never get to observe: its end. In a paper first presented to the Academy of Management in 1992 under the title "Reversal of a Successful Gainsharing Program: Experimental Evidence of Residual Effects", and published in fuller form in Group & Organization Management in 1994, they measured employees' attitudes, behaviours and cognitions during the plan and after the group bonus was eliminated.47
Three months after the bonus ended, the group that had been under the plan showed higher moral commitment, more prosocial behaviour and less intention to leave than the control group.
The proceedings title adds a word to what the 1994 abstract says. It calls the programme successful. The plan was not ended because it failed. The abstracts do not give the reason it was ended, and this chapter does not supply one. What they make possible is a reading of what a plan leaves behind when the money stops for reasons unconnected with whether it worked: a change in ownership, a budget decision, a new management. Those reasons are common. Chapter 5 read Dong-One Kim's finding that sound financial health and capital investment were associated with a plan's survival, which implies that plans also end for reasons that have little to do with how well they work.
The residue the study found is relational. Moral commitment is the sense that one ought to stay with the group. Prosocial behaviour is help offered beyond what the job requires. Lower intention to leave is a preference to remain among these people. None of these is a productivity figure. All three are what one would expect if the plan's months of increased job talk and idea exchange had built something between people, and if that something was not dissolved the moment the cheque stopped.
The limits are plain. One firm. Three months after the end. Attitudes and behaviours measured, not long-run output. The study cannot say whether the residue lasted a year, or whether a firm that ended a plan clumsily, with a broken promise rather than an explanation, would see the opposite. Chapter 7 reads a case of exactly that kind. But the direction of the finding is the direction this book has argued from the start. If a plan's first return is talk, and talk builds relationship, then the relationship is the part that can outlive the money.
Santa Anna's book, in its chapter on policy and governance, makes an observation about citizens and their institutions that transfers to the people in a delivery firm almost unchanged:
Trust, it turns out, is extraordinarily efficient. Suspicion is extraordinarily expensive.48
The study does not measure trust directly. But commitment, helping and staying are what trust looks like from the outside, and a firm that keeps them after a plan ends has kept something it would otherwise have had to buy.
The third study asks a question the first two take for granted. If a gainsharing plan works partly because people pay more attention to one another's work, how do managers respond to that attention?
The free-rider objection, read in Chapter 3, assumes that group rewards fail unless colleagues watch each other. Eugene Kandel and Edward Lazear's answer to it was peer pressure and mutual monitoring. Welbourne and Ferrante point out, in the opening of their 2008 study, that this answer contains an untested assumption: that employees will in fact monitor one another, and that their managers will support them when they do.49
They tested it with 203 employees in a firm providing customs brokerage and freight-forwarding services, under two pay conditions: traditional merit pay, and merit pay combined with gainsharing. They distinguished two kinds of peer monitoring. Observational monitoring is noticing what one's co-workers are doing. Advisory monitoring is reacting to it: saying something, offering a correction, suggesting another way.
Observational monitoring was positively associated with managers' performance ratings of the worker under both pay conditions. Managers liked people who paid attention. Advisory monitoring was different. Under gainsharing it was positively related to the manager's rating of the worker. Under traditional merit pay it was negatively related.
Read that slowly, because it is one of the most practically useful findings in this book. The same behaviour, an employee telling a colleague that something could be done better, was rewarded by managers in one pay system and penalised in the other. Under merit pay, the employee who speaks up about a colleague's work looks, to a manager, like someone who is overstepping, or perhaps someone who is competing. Under gainsharing, the same employee looks like someone looking after a shared result.
What the study shows is an association in one firm, measured through managers' ratings. It cannot show that the pay system caused managers to rate differently; other differences between the groups may have contributed, and the abstract does not say how the two conditions were assigned. It cannot tell us whether the employees themselves felt the same difference. But it reaches directly into the place Chapter 3 identified as the plan's most fragile joint: the supervisor. A supervisor's rating is one of the clearest signals an employee receives about which behaviours are safe. If advisory monitoring is rated down, people stop advising, and the up-current narrows to the remarks that offend nobody.
Santa Anna's book puts the general principle without qualification:
safety produces better outcomes than fear.50
Welbourne and Ferrante's result shows what that principle looks like at the level of a single performance review. The safety to tell a colleague something useful is not only a matter of culture. It is written into how a manager scores the person who said it. A leader who introduces a shared measure and leaves the rating system unchanged has asked people to speak up and then kept a scoring rule that, in at least one firm's evidence, punished them for it.
The finer points for a service leader
Haute Lumière · The Gain Was Always Shared · Chapter Six
The three studies, read together, give a leader in service work six points of practice. Each is drawn from what the studies found, and each states what they cannot show.
Measure the talk. Hanlon and Taylor measured communication behaviours because they thought it was where the plan acted first. A service firm can measure something similar without a research team: how often ideas are raised in team meetings, how many cross-team suggestions arrive, how often people ask a colleague in another role for help. These are noisy counts. They are also the earliest available signal that a plan is doing anything, months before the shared measure has had time to move.
Expect the first six months to show in conversation, not in the measure. The service study found its effects in the early stage of the plan. A leader who looks only at the outcome measure in that period, and sees little movement, may conclude the plan is failing while the up-current is just beginning to run. Write down before the plan starts what the first six months are for: more talk about the work, more ideas moving between people. Judge them on that.
Change the rating before you launch the plan. If supervisors' ratings still reward individual output and penalise people who comment on others' work, the Welbourne and Ferrante result suggests the plan will be fighting its own appraisal system. Add one line to every supervisor's review of their staff: whether the person helped a colleague improve, and how. Add one line to every supervisor's own review: whether the people who spoke up in their team were rated fairly for it.
Teach the difference between noticing and advising. Observational monitoring was rated well under both systems. The harder behaviour, and the one that carries the value, is advisory: saying something. People need language for it that does not sound like an accusation. A short shared form helps: I noticed this; here is what I have seen work; what do you think? It is small, and it is the difference between a team that watches and a team that improves.
Plan for the end as well as the start. The residual-effects study suggests that some of what a plan builds can survive the end of its bonus. A leader can make that more likely by deciding in advance which parts of the participation system are not tied to the money: the meeting, the answer rule, the visible measure, the habit of naming who found what. If those are written into the way the team works, rather than into the bonus document, they do not end when the bonus does.
Do not read service evidence as a forecast. Two firms, three studies, results stated in abstracts. They are the right direction and the wrong size of evidence for a prediction. A service leader should treat them as Chapter 3 suggested treating every figure in this literature: as reasons to run their own plan as a study, with a before, a comparison and a written expectation.
What follows is a composite, built for teaching. It is not a real organisation, its details were set by the writer, and it describes the participation system only, not any reward design.
A physiotherapy practice of thirty people across two sites has a problem that every person on the front desk can describe and nobody has fixed. Patients arrive without the referral letter their insurer needs. The front desk spends part of every morning on the telephone. Clinicians start late. Patients wait.
The practice manager introduces one practice before anything else. Every Tuesday, the two front desks and one clinician from each site hold a fifteen-minute call. Each person brings one thing they noticed about the week. The manager's only role is to write the notes and answer, within a week, any idea that needs her decision.
For the first month the calls are mostly complaints. In the second month the receptionist at the smaller site says that her colleague at the larger site has a way of reminding patients about letters by text the day before, and asks how she does it. The larger site's receptionist explains. The smaller site adopts it. Nobody had told anyone to share it. It had simply never been anybody's business to ask.
In the third month the clinician at the larger site says, carefully, that the front desk at her site books new patients into the first slot of the day, which is the slot most likely to be lost to a missing letter, and wonders whether they could go later. It is the advisory kind of remark: a comment on someone else's work. The manager notes it and, in the receptionist's next review, records it as a contribution from both of them, the clinician for raising it and the receptionist for trying it. Late starts fall over the following weeks. The practice posts the late-start count on the staff-room wall.
The composite is not evidence. It is an illustration of the order this chapter recommends for service work: talk before money, a rating that rewards the advisory remark, and a visible count that lets people see their conversation change something.
Chapter 5 worked through factory evidence and found the gains concentrated in quality and downtime. This chapter adds the mechanism by which, in service work at least, those gains are likely to arrive: people talking more about the job, passing more ideas to one another, and feeling safe to comment on each other's work when the pay system and the rating system say that is what a good colleague does.
It also adds something to the living-systems reading. Santa Anna's book lists, among the vitality-centred workplace's design principles, one about authority:
When authority is distributed to the people closest to the work, creativity and ownership become possible51
The service studies show a modest, concrete version of that. The authority distributed is not the power to hire or to set budgets. It is the standing to say to a colleague, I think there is a better way, and to be rated well for it. It is the smallest unit of distributed authority there is, and it may be the one from which the larger ones grow.
The book's sister volume on gift and reciprocity, gift and reciprocity, reads the laboratory evidence on why an act of attention tends to be returned. The service studies are a field version of the same pattern. A plan that changes what people say to each other gives them something to return.
Use this for the first six months of any participation system in service work, before the shared measure has had time to move. Keep one page per month. Every entry is a count, a date or a quotation, not an opinion.
Header
Team or unit: · People covered: ___ · Month ___ of 6 · Comparison team (if any, and why it is comparable):
1 · Talk about the work
2 · Ideas moving between people
One example, in the words of the person who adopted it:
3 · Noticing and advising
The shared phrase this team uses for an advisory remark:
4 · The rating
Any case this month where someone who made an advisory remark was rated down for it: yes / no · what was done:
5 · The measure (watched, not yet judged)
Outcome measure chosen: · this month: ___ · same month last year: ___
Written expectation for month six, set before month one:
6 · What survives
Parts of the participation system written into how the team works, independent of any bonus: meeting / answer rule / visible measure / naming who found what / other:
If the reward ended tomorrow, which of these would continue?
Chapter Seven
Promises, Prices and the Operating Room
Haute Lumière · The Gain Was Always Shared · Chapter Seven
The cases in Chapters 5 and 6 are, on the whole, encouraging. A plant that left piecework and saw defects fall. A Scanlon plant watched for up to nine years. A delivery firm whose people began to talk more about the job. They are the right cases to read first, because they show what the idea looks like when it works.
This chapter reads two rooms where it is harder. The first is a firm whose Scanlon plan ran into trouble, studied closely by an ethnographer who stayed long enough to see why, and read beside Denis Collins's work on plans that failed. The second is the cardiac catheterisation laboratory, where a set of American hospitals shared savings with the physicians who worked in them, and where the most careful study found that the gain came mostly from somewhere other than where one might expect.
Each room teaches one thing the earlier cases could not. The first teaches what happens when a plan is used as a bargaining chip and a promise is broken. The second teaches the question every leader should ask before a single gain is shared: where, exactly, did it come from? Both end in the same place. A shared gain is honest only when the people sharing it can see its source, and when the thing the organisation exists to protect has been protected on the way.
In 1991 Kenneth Thornicroft published a qualitative, ethnographic study in Employee Relations with a title that names its subject in nine words: "Promises Kept, Promises Broken: Reciprocity and the Scanlon Plan".52 It investigates the particular problems one American firm experienced when it implemented a Scanlon productivity gainsharing plan.
The abstract states two conclusions, and they are worth reading as the author wrote them rather than paraphrased into something stronger.
The first is that such plans should not be treated as simple quid pro quos for collective bargaining concessions, because they necessarily pose significant impediments to the maintenance of the status quo. The phrase carries a whole history. A firm under pressure asks its union for concessions, on wages or work rules, and offers a gainsharing plan in exchange: give up something certain now, and share in what you help us save. On paper the trade is clean. Thornicroft's argument is that it misunderstands what the plan is. A Scanlon plan is not a payment. It is a change in how decisions are made, and a firm that installs it as a payment, intending to keep everything else as it was, has bought a change it does not intend to make.
The second conclusion is that successful implementation requires both a sound programme structure, especially concerning the bonus formula, and a change in the organisation's culture whereby all employees are able to participate meaningfully in decision-making.
The word both carries the weight. Thornicroft does not say that the formula is unimportant; he names it specifically. He says that a sound formula is not sufficient. The participation has to be real.
The abstract does not tell us the firm's name, its industry, the content of the promises or exactly which were broken. This chapter does not supply them. What the title supplies is the mechanism. The study is about reciprocity, and its title divides promises into two kinds. Some were kept. Some were broken. An ethnographer who frames a study that way has found that what decided the plan's course was not the arithmetic but whether the people in it believed the other side would do what it said.
Chapter 3 introduced Denis Collins's work: a 1996 article on the lessons of one gainsharing plan's death, and a 1998 book, Gainsharing and Power, built on close study of six Scanlon-type plans.53 Read beside Thornicroft, the two bodies of work describe the same failure from two angles.
Collins's angle is power. The tenth chapter of his book is titled "Power Games, Outcomes, and Lessons Learned of Scanlon-Type Gainsharing Plans". The common reading of his work, which Chapter 3 followed, is that plans fail less often because the formula is wrong than because participation is never real: managers keep the right to decline ideas without reasons, committees become places where managers report and employees listen, and supervisors experience the plan as an audit.
Thornicroft's angle is promise. A firm that offers a plan in exchange for concessions has made a promise about the future: that ideas will be heard, that gains will be returned, that the new arrangement means what it says. When the culture does not change, the promise is broken not by a single act but by a thousand small ones. An idea declined without a reason. A committee meeting cancelled. A base raised after a good year. Each is minor. Together they tell the people who gave up something certain that what they received in exchange was not real.
Ammanuel Santa Anna's Ecstatic Economics has a sentence about governance that names the felt experience on the far side of a broken promise. It is written about citizens and the state, and it describes a workforce after a plan has gone wrong with uncomfortable precision:
a felt sense of being governed at rather than governing together54
A plan offered as a trade and run without real participation leaves people governed at. The committee exists, the formula exists, and the people on the floor experience both as things done to them. Collins called the result power games. Thornicroft called it a broken promise. The book calls it a felt sense, and it is right that it is felt before it is analysed. People know the plan is hollow long before any survey asks them.
What the first room asks of a leader
Haute Lumière · The Gain Was Always Shared · Chapter Seven
The case is a warning, but a warning with a shape, and the shape gives a leader four specific practices.
Do not offer a plan as the price of a concession. If a firm needs concessions, ask for them as concessions, with the reasons, and negotiate them on their merits. If it wants a participation system, build one because it wants one. Bundling the two tells everyone that the plan is a sweetener, and it puts the plan's credibility at the mercy of the bargaining relationship at its worst moment. Where the two genuinely must come together, as they did in the failing mill of Chapter 2, say plainly what is changing in how decisions are made, and do it before the first payment, so that people can see the change is real.
Write the promises down, and keep a ledger of them. Thornicroft's title implies that the people in his firm were counting. They always are. A firm can make the counting visible and honest by listing every promise the plan makes, about answers, about the measure, about the base, about jobs, and recording each month whether it was kept. A promise that has to be broken, for a real reason, should be broken in the open, with the reason, by the person who made it.
Treat the formula and the culture as one design problem. Thornicroft named both. A firm that spends six months on the formula and an afternoon on participation has decided which one it believes in, and its people will read the decision correctly.
Keep a record of what fails. Collins's 1996 article exists because someone stood by a plan's grave long enough to write fifteen lessons. Thornicroft's study exists because someone stayed in a troubled firm long enough to see why. A leader can do the same thing internally, and should.
The second case comes from a field where the word gainsharing means something rather different, and where the difference is instructive.
In American healthcare, gainsharing has come to describe arrangements in which a hospital shares with physicians some of the savings that result when costs fall. The physicians are often not employees of the hospital, which is part of why the arrangements were legally sensitive. In 2008 Jonathan Ketcham and Michael Furukawa published a study in Health Affairs of thirteen such programmes, approved by the Office of Inspector General of the Department of Health and Human Services, and their effects on patients receiving coronary stents.55
Their abstract reports four findings.
Compared with other hospitals, the gainsharing hospitals reduced costs by seven point four per cent per patient.
Of those savings, ninety-one per cent came from lower prices and nine per cent from lower utilisation.
The available measures of access and quality suggested that neither was reduced.
Nor was access to drug-eluting stents reduced, before 2006.
Each of these findings needs careful reading, and the second needs the most.
A reader who arrives at this case from the mill and the Scanlon plant brings an expectation. Gainsharing, in the industrial tradition, rewards people for finding better ways to work: fewer defects, less downtime, less waste. In a catheterisation laboratory one might expect the equivalent to be fewer devices used per procedure, fewer unnecessary tests, a smoother sequence in the room.
The study found that about nine-tenths of the savings came from lower prices. Only about one-tenth came from using less.
The abstract does not explain the mechanism behind the lower prices, and this chapter does not supply one. What can be said is what the finding means for anyone designing a shared-savings arrangement. A cost measure moves for two quite different reasons: because people change what they do, and because the organisation pays less for what it buys. Both are real savings. They are not the same kind of gain, and they do not rest on the same knowledge.
Chapter 2 asked of any shared measure: if the people in this room changed how they work, would this number move, and would they be able to see why? Chapter 3 read the fourth critique of gainsharing, that a measure moves for reasons nobody inside the group controls. The cardiology case adds a refinement. A measure can move because of something the group does influence, such as which devices are chosen, which can in turn change what the organisation pays, without anyone in the room working differently in the clinical sense. That is neither illegitimate nor the same thing as the night-shift operator's idea about a sealing jaw. A leader who wants to know what the organisation is actually rewarding has to separate the two.
Santa Anna's book, writing about wellbeing measures that governments adopted in place of growth figures alone, gives a sentence that holds at the scale of one shared measure:
The framework is necessary but not sufficient.56
A cost measure is a necessary frame. It is not sufficient to show what kind of improvement occurred. The study shows the value of taking a measure apart: not only did costs fall, but how much of the fall came from price and how much from practice. That decomposition is what lets the people inside an arrangement see what they are sharing.
The guardrail
Haute Lumière · The Gain Was Always Shared · Chapter Seven
The other half of the cardiology finding is the half that should be read first by any leader in a field where the product is care.
The obvious risk of rewarding cost reduction in medicine is that it reduces care. A physician rewarded for using fewer or cheaper devices has, at least in principle, an incentive to use fewer or cheaper devices than a patient needs. The Office of Inspector General's involvement reflects exactly that concern. Ketcham and Furukawa's abstract states the concern directly before reporting the finding: that while gainsharing might reduce costs by aligning hospital and physician incentives, concerns remained about quality and access.
Their finding is that the available measures of access and quality suggest neither was reduced. The word available is the authors' and it is honest. The study could look only at the quality and access measures the data allowed. It cannot show that no patient received less than they should have, and it cannot show that the same result would hold in another specialty, another country or another regulatory setting. It also does not say how the arrangements themselves were designed, and this book does not describe that design.
What the case shows is a pattern: savings were shared, and the measures available did not show the harm that was feared. A leader can take from it a principle that applies far beyond cardiology. In any organisation whose purpose is something other than cost, a shared measure of cost or productivity needs a guardrail on the purpose, measured alongside it, visible to the same people, and given priority over it. In a hospital, the guardrail is quality and access. In a school, it is learning. In a firm making food, it is safety. In a software team, it is the reliability of what the customer depends on.
Santa Anna's book, describing the felt difference between two kinds of governance, puts the distinction a guardrail exists to protect:
It is the difference between being managed and being cared for57
A patient in a hospital where the shared measure is cost, and nothing else, is being managed. A patient in a hospital where cost is shared only inside a guardrail on quality and access has some assurance of being cared for. The guardrail is what tells the people working in the room, and the people being treated in it, which of the two the organisation means.
The broken promise and the catheterisation laboratory seem to have little to do with one another. One is an ethnography of a troubled plan in an unnamed firm. The other is a quantitative study of thirteen programmes in a highly regulated field. Read together, they describe the two ways a shared gain can stop being honest.
The first is a gain whose promise is not kept. People are told they will share in what they help produce, and then the participation that would let them produce it is withheld, or the rules change after a good year, or ideas go unanswered. The gain may still be measured. It is no longer shared in any sense its participants recognise.
The second is a gain whose source is not visible. A measure moves, savings are shared, and nobody has separated how much of the movement came from changed practice, how much from changed prices, and how much from things nobody inside controlled. The sharing may be perfectly fair. Without the decomposition, nobody can tell.
The way through in both rooms is the same practice the earlier chapters arrived at from other directions. Make the promises explicit and keep them in the open. Take the measure apart so people can see where the gain came from. Protect the purpose with a guardrail that comes before the gain. And let the people covered take part in deciding all three.
Santa Anna's book has a sentence, written about the relation between individual change and structural change, that names why both rooms failed or nearly failed in the same way:
Change the policies without changing the somatic pattern, and the old economics will reassert itself through a thousand unconscious decisions.58
Thornicroft's firm installed a policy, a Scanlon plan, without changing the pattern of who decided things, and the old pattern reasserted itself through the small broken promises his title points to. A shared-savings arrangement without a guardrail risks the same reassertion in a different form: the old logic that cost is the only thing that counts, returning through a thousand clinical choices. The guardrail and the kept promise are how the new pattern is held in place long enough to become habit.
Name the purpose first, and measure it beside the gain. Before any shared measure is chosen, write down what the organisation exists to protect and how it will be measured. Put that measure on the same board as the gain measure, above it.
Agree in advance that the guardrail wins. If the quality or access measure falls, the gain measure does not count, whatever it shows. Say so in writing before the arrangement begins. This is a rule about precedence, not about how any gain would be divided, and it can be decided without any discussion of reward design.
Decompose every gain. Each quarter, separate the movement in the measure into what came from changed practice, what came from changed prices, and what came from events outside the group's control. Show the three to the people covered.
Watch the legal setting. In healthcare, and in other regulated fields, shared-savings arrangements can raise legal questions that do not arise in a factory. The cardiology programmes in this study were approved by a regulator. A leader in such a field should take qualified legal advice before designing anything, and this book offers none.
Never trade the plan for a concession. Thornicroft's warning applies with special force where professionals are not employees, and where trust between the organisation and the people in it is already strained.
Use this before any shared-gain arrangement begins, and monthly once it is running. It has two halves: the promises the arrangement makes, and the sources of whatever gain it records.
Header
Organisation or unit: · People covered: ___ · Month ___ · Prepared by:
1 · The purpose and its guardrail
What this organisation exists to protect, in one sentence:
The guardrail measure(s): · this month: ___ · baseline: ___
Written rule, agreed before the start: if the guardrail falls, the gain measure does not count. Agreed on (date): · by whom:
2 · The promises
| Promise made | To whom | Kept this month? yes / no | If no, the reason given, and by whom |
|---|---|---|---|
| Every idea answered in writing within ___ days | |||
| The measure visible to everyone covered | |||
| The rules for changing the base unchanged without notice and a vote | |||
| No role removed by an implemented idea without (stated alternative) | |||
| Other: ______ |
Was the arrangement offered in exchange for a concession? yes / no · If yes, how the participation was made real before the first payment:
3 · The sources of the gain
Total movement in the gain measure this period:
Of which, from changed practice (what people did differently):
Of which, from changed prices (what the organisation paid):
Of which, from events outside the group's control:
Shown to the people covered? yes / no · how:
4 · Power
Who decided the most consequential change this month, and were the people covered part of the decision?
5 · Advice
Legal or regulatory review completed (where the field requires it): yes / no / not applicable · date:
6 · The record
The most important promise kept this month:
Any promise broken, and what was done about it:
Chapter Eight
The Field Manual
Haute Lumière · The Gain Was Always Shared · Chapter Eight
The chapters before this one read gainsharing from four directions: its origin in a mill that asked first (Chapter 2), the five ways it fails (Chapter 3), the living-systems reading of it as a circulation (Chapter 4), and the published cases, figure by figure, with a composite packaging plant laid out month by month (Chapter 5). This chapter turns all of that into a sequence a leader can run.
It is written for the person who will actually do the work: a plant manager, a hospital department head, the owner of a firm of eighty, the director of a public agency. It covers roughly the first eighteen months, in phases, with the weeks marked where the timing matters and left loose where it does not. For each phase it names who is involved, what is decided, and who decides it. It ends with a single page that holds the whole sequence.
Three things should be said before the first week.
First, this is a sequence for building a participation system, and it treats a shared reward as one late stage of that system. Where the sequence reaches the design of a reward, it says only that the design is the organisation's own, made with the people it would cover and put to their vote. This book does not describe how a gain is calculated or divided. That is not modesty. The research read in Chapters 3 and 5 is consistent on the point that the design of the formula explains less about whether a plan lives than the participation around it, and the formula is the part most specific to each firm's accounts. A manual that told you how to divide a gain without knowing your books would be telling you the one thing it could not know.
Second, the timings are a recommendation drawn from the studies cited, not a finding of any one of them. Some organisations will move faster; most will find that one phase needs repeating. The order matters more than the pace.
Third, the sequence assumes the leader means it. Ammanuel Santa Anna's Ecstatic Economics states the condition for any practical framework in two short sentences that serve as the epigraph for this whole chapter:
A theory that cannot be practiced is a poem. A practice without theory is a reflex.59
The previous chapters were the theory. What follows is the practice. Neither stands without the other.
Before the phases, the people. A participation system has seven roles. In a small organisation one person may hold two of them; in a large one, a role may be held by several people. What matters is that each is named, in writing, before the work begins.
The sponsor. The person with the authority to commit the organisation for at least two years: an owner, a chief executive, a board, a director-general. The sponsor's job is not to run the system. It is to protect it: to keep it funded through a bad quarter, to stop it being cancelled by someone who arrives later, and to say in public, at the start, that it will be given time. Dong-One Kim's survey of 211 organisations with gainsharing experience associated survival with sound financial health and major capital investment alongside the plan.60 Both are sponsor decisions. A sponsor who will not commit for two years should not start.
The unit leader. The person who runs the plant, department, clinic or office the system covers. The unit leader answers ideas the committees cannot decide, chairs the screening committee in its first months, and speaks first at the weekly meeting. This is the role that costs the most time, and the time cannot be delegated in the first ninety days, because the people covered are watching whether the person in charge actually reads what they write.
The supervisors. The people who manage the daily work of teams. Chapter 3 named the supervisor as the person a participation system most often threatens, because ideas arriving at a committee make visible what a supervisor knew and did not fix. In this manual the supervisors are briefed before anyone else, their role is redesigned before the system starts, and they are measured on one thing in the first phase: the share of their team's ideas answered on time.
The committees. A production committee for each team, shift or area, and a screening committee above them. Members from the floor are elected, not appointed; each production committee includes its supervisor; the screening committee includes the unit leader, the finance person and elected members. Committees are formed in the middle of the sequence, not at the start. The answer rule comes first, so that the committees inherit a habit rather than having to invent one.
The secretary. One person, elected from the floor, who keeps the record: every idea received, the date, the person, the answer, the date of the answer, whether it was implemented, and what happened to the measure afterwards. The secretary is the memory of the system. Chapter 5 described a firm becoming its own case study; the secretary's ledger is that study. The role should carry protected time, at least two hours a week, and the secretary should be someone the floor trusts to record a declined idea as faithfully as an adopted one.
The finance person. The person who can open the books far enough for the measure to be believed. Chapter 2 argued that both the Scanlon and Rucker plans required the firm to show its people a number they could check, and that a shared measure people cannot inspect is a promise rather than a measure. The finance person's job is to make the chosen number inspectable: where it comes from, how it is calculated, what moves it that the unit does not control. They sit on the screening committee and explain the number, in plain words, every month.
The union, where there is one. Kim's survey associated union participation in running a plan with its survival, and Jeffrey Arthur and Dong-One Kim's comparison of two plants found employees more willing to share high-risk knowledge where labour and management cooperated.6061 Where a union represents the people covered, it is invited in before the supervisors are briefed, not after the plan is announced, and it holds seats on the screening committee by right. Where there is no union, the elected committee members carry the same function, and the sequence should not be used to argue that a union is unnecessary.
One further role is temporary. Outside help, if it is used, is a facilitator who builds the committees and trains them, then leaves on a date set in advance. Chapter 5 resolved the disagreement between Robert Bullock and Mark Tubbs, who found outside practitioners associated with success, and Kim, who found reliance on outside consultants associated with failure to survive: use outside help to build the system, never to operate it.6260 The leaving date is written into the facilitator's terms on the first day.
What happens. The sponsor and unit leader decide whether to begin. They do not announce anything.
Week one. The unit leader runs the Origin Reading from Chapter 2 with two people who do the work. One hour. The four questions: what do the people doing this work know that management does not; how is an idea answered today; is there a number this group could move and see; and is the group drawn around cooperation or around the organisation chart. The answers are written down. They become the baseline every later review returns to.
Week two. The finance person and the unit leader look at the last two years of whatever operational data exists: defects, rework, downtime, errors returned by the next team, waiting times, complaints, overtime, grievances, absence, voluntary leavers. They do not choose a measure yet. They find out what is already counted, how reliably, and how far back. A firm with eighteen months of clean monthly data has a baseline; a firm with none must start collecting now, because Chapter 9 will show that a change cannot be seen without a before.
Week three. The Five-Critique Pre-mortem from Chapter 3, run by the unit leader with one supervisor and two people the system would cover. Imagine the system abandoned two years from now; write the most likely reason under each critique; write the step that would prevent it. Any critique without a preventing step is a reason to wait.
Week four. The sponsor's decision. Four questions, answered in writing. Will the organisation commit to the system for at least two years, including through a weak quarter? Can it make a credible promise about what happens to a job that an idea removes? Is the unit financially sound enough that the system will not be read as a rescue that becomes a cut? Is there any obvious inequity in pay or treatment within the unit that a visible measure would expose?
That last question matters more than it looks. Santa Anna's book, in its discussion of pay transparency, warns that an unjust system made visible produces suspicion rather than trust, and Chapter 3's critiques and Chapter 4's reading both point the same way. A unit with a known inequity either fixes it first or names it at the start with a date for fixing it. Starting silently over it is the one choice that reliably fails.
Who decides. The sponsor decides whether to begin. The unit leader decides whether the unit is ready. If either says no, the work for the next ninety days is whatever made them say no.
What can go wrong here. Beginning because a competitor has, or because a consultant has proposed it, or because the unit is in crisis and something must be seen to happen. Chapter 2 noted that crisis can open the door to participation and cannot keep it open. A system begun in panic has to find a reason to continue once the panic passes. Write that reason down in week four.
What happens. The people most able to make the system work, and most able to quietly stop it, are brought in before anyone else.
Week five. Where there is a union, the unit leader and sponsor meet its officers. They bring the Origin Reading, the pre-mortem and the sponsor's four answers. They say what they intend: a rule that every idea will be answered, a measure made visible, committees with elected members, and only later, if the first three work, a shared reward designed with the people it covers and put to their vote. They ask what the union would need for this to be something it helps run rather than something done near it. They write the answer down.
Weeks five and six. The unit leader meets each supervisor alone. One conversation each, forty-five minutes, with three parts.
The first part is the honest description of what will change. Ideas will arrive at a committee without passing only through the supervisor. Some of them will describe problems the supervisor knew about.
The second part is the new role. The supervisor becomes the person who brings the team's ideas forward and is named when they are adopted. In the first ninety days the supervisor is measured on one thing: the share of the team's ideas answered on time. Not their content, not their volume, not the defect rate.
The third part is a question: what would you need for that to be true? Time, training, cover for the shift while you attend committee, a guarantee that an old problem surfaced by a card will not be held against you. Write down each answer and give a reply to each within a week. The supervisor is about to be asked to trust a promise that every idea gets an answer. Answer theirs first.
Santa Anna's book names the principle under this phase in three words, at the head of its design principles for the workplace:
Vitality requires agency.63
The supervisor's agency is the first to be threatened by a participation system and the first to be rebuilt. A supervisor who has been given a part to play and the means to play it is an ally. A supervisor who has been told the system exists is a closing valve.
Weeks seven and eight. All supervisors together, with union officers where present. The unit leader presents the answer rule in draft, the reporting line for ideas, and the supervisors' measure. The draft is changed by what they say. It is then fixed.
Who decides. The unit leader decides the supervisors' role and measure, after hearing them. The union decides its own participation.
Phase three · The answer rule (weeks nine to twenty-two)
Haute Lumière · The Gain Was Always Shared · Chapter Eight
What happens. One rule is announced to everyone, and nothing else. This is the longest phase and the most important. Everything after it depends on the habit it builds.
Week nine: the announcement. The unit leader speaks to everyone covered, in person, by shift. Five things are said and nothing more.
The last sentence costs nothing to say and removes the suspicion that the answer rule is a preliminary to a target. People who suspect a target hold back their best ideas until they see it.
Weeks nine to twenty-two: answering. The secretary is elected in week nine, from the floor, by whatever simple method the unit already uses. From the first card, the secretary records every idea and every answer.
The unit leader reads every answer before it is posted, for the first eight weeks. Not to change it. To learn what the people covered are noticing, and to see which supervisors decline without reasons. A declined idea with a reason is a good answer. "No, because the line is being replaced in March" tells the person they were heard and teaches them something about the system. A decline with no reason is the commonest way an answer rule dies.
At the weekly meeting, the unit leader speaks first, and speaks about something uncertain. Chapter 10 returns to this habit and Chapter 12 reads it as the beginning of the circulation; here the practical point is that a leader who wants ideas to move upward must show, every week, that saying something unfinished is safe.
Week fourteen: the first count. The secretary reports four numbers to everyone covered: ideas received, from how many different people, the share answered on time, and the share implemented. Post them by the board. These are the vital signs Chapter 4 named for the up-current and the answer. In the composite packaging plant of Chapter 5, ideas rose from eleven in the first month to forty in the third; a real unit may see more or fewer. What matters is the direction, and whether the number of different people writing is rising, which shows the channel is being trusted beyond the few who would have spoken anyway.
Week twenty-two: the readiness review. The sponsor, unit leader, supervisors, secretary and union meet for two hours. Three questions. Is the answer rule being kept: is the on-time share above nine in ten? Is the number of people writing rising? Have any supervisors stopped answering? If the answer to the first is no, phase three continues until it is yes. Do not proceed to the measure while ideas wait unanswered; a visible number posted beside unanswered cards tells everyone the organisation values the number more than them.
Who decides. The unit leader decides when the answer rule is being kept. The secretary's record is the evidence, and nobody overrules the record.
What happens. A number the unit can move by changing how it works is chosen, made inspectable and posted weekly. No money is attached to it.
Weeks twenty-three to twenty-six: choosing. A small group chooses the measure: the finance person, one supervisor, two people from the floor chosen by the others, and the secretary. They start from the operational data assembled in week two and the ideas received in phase three, which show where the people covered see waste and error. They apply the test from Chapter 2: if the people in this unit changed how they work, would this number move, and would they be able to see why?
Kaufman's study of 112 Improshare firms and the Hatcher and Ross plant both located much of the gain in fewer defects and less downtime.65 Those are good candidates because the people doing the work can usually see what causes them. A financial ratio may come later. In this phase choose something close to the work and countable weekly: defects per thousand, rework hours, minutes of downtime, errors returned, waiting time, complaints.
The group also writes down, in advance, the things outside the unit's control that move the measure, and how each will be noted when it happens. A supplier change. A new machine bought by head office. A change in case mix. A contract won or lost elsewhere. Writing these down now is the defence against the fourth critique of Chapter 3, the number that moves with the weather.
Week twenty-seven: opening the books. The finance person explains the measure to everyone covered: where it comes from, how it is counted, who counts it, what it was last year. Anyone can ask to see the underlying record. Santa Anna's book, writing about pay transparency, puts the effect of this step in a single sentence:
The nervous system relaxes when it can see the landscape clearly.66
Weeks twenty-seven onwards, the measure is posted weekly by the board where the answers go, broken down by team or line where it can be. When an implemented idea is connected to a movement in the number, the name of the person who had it is written beside the line. Chapter 4 called this the down-current reaching the person, not only the group.
Weeks twenty-seven to forty: collecting. Collect for at least three months before anything further. Six is better; twelve, counted together with whatever historical data exists, is better again. Chapter 9 explains why. A change can only be seen against a before, and a before collected deliberately, with the outside events noted, is worth more than any later argument about what the number would have been.
Week forty: the expectation. Before the committees are formed, the unit leader, secretary and finance person write down what they expect to change over the next year, by how much and by when: ideas per month, median days to answer, share implemented, and the measure. They seal it, in the sense that it is dated and filed where it cannot be revised. At the first review it is opened and compared with what happened. This one page is the most useful defence against the survivorship problem Chapter 3 described, because it prevents a firm from deciding afterwards what it had hoped for.
Who decides. The measurement group chooses the measure; the finance person decides how it is counted; nobody may change either without saying so on the board.
What happens. The answering habit passes from individuals to elected bodies, with help to build them if help is wanted.
Week forty-one: elections. Each team, shift or area elects members to its production committee. Two or three from the floor plus the supervisor is a workable size. The screening committee is formed from elected members of each production committee, the unit leader, the finance person, the secretary and, where present, union officers. The union may nominate its own seats; it should not be asked to compete for elected ones.
Weeks forty-one to fifty-two: building, with a leaving date. If outside help is used, this is when. A facilitator trains the committees to read the measure, to evaluate an idea against cost, safety and quality, to write a declined answer with a reason, and to run a meeting in which people from the floor speak as much as managers. The facilitator's leaving date was set in their terms; by week fifty-two the committees run themselves. The facilitator's last task is to watch a meeting they do not chair and say afterwards, privately, to the unit leader, what they saw.
The handover of the answer rule. From week forty-four, production committees answer ideas within the team's own work, and the screening committee answers anything wider. The unit leader answers only what the screening committee cannot decide, and says why in the answer. The ten-day rule does not lengthen. A committee that cannot meet often enough to keep it meets more often or delegates to two members between meetings.
What committees decide. Which ideas are implemented within an agreed budget limit per idea, without further approval. The limit is set by the sponsor in week forty-one and posted. Ideas above it go to the unit leader with the committee's recommendation. A committee that can implement nothing is a reporting meeting, and Chapter 3 described where those end: in the common reading of Denis Collins's six plans, plans fail less often because the formula is wrong than because participation is never real.67
The second-order ideas. Jeffrey Arthur and Lynda Aiman-Smith's study of four years of suggestions at one plant proposed that a maturing plan should see ideas move from fixes within a process towards ideas that question the process itself, and Arthur and Kim's comparison of two plants found that willingness to offer that riskier kind of knowledge depended on trust between labour and management.6861 The committees should count both kinds. The secretary can mark each idea as a fix or as a question about the process. If the second kind does not grow over the year, the committee asks why, and the usual answer is that it is not yet safe to say a manager's design was wrong.
Santa Anna's book names the condition in its list of workplace design principles:
Vitality requires information.63
The passage goes on to describe organisations that suppress feedback as losing the ability to adapt and self-correct. A committee that receives only safe ideas is receiving less information than the unit holds. The remedy is not to exhort people to be braver. It is for the unit leader to answer the first process-questioning idea with visible thanks, whatever its fate, and for the committee to implement one such idea quickly enough that everyone sees it.
Week fifty-two: the first annual review. Open the sealed expectation. Compare it line by line with the secretary's record. Write down, plainly, what was expected and did not happen, and what happened that nobody expected. Post the comparison. This is the first page of the organisation's own case study.
Who decides. The people covered decide who sits on the committees. The sponsor decides the budget limit. The committees decide within it.
What happens. If, and only if, the first year has shown the circulation working, the people covered are invited to help design a shared reward and to decide by vote whether to adopt it.
The gate. Four conditions, checked at the first annual review. The answer rule kept at nine in ten or better for six months. The number of different people contributing ideas steady or rising. The measure visible weekly for at least six months, with outside events noted. Committees meeting as scheduled and implementing ideas within their limit. If any condition is not met, the design waits. The work is whatever condition is unmet. Chapter 2's test applies: a firm that has not learned to ask and answer will find that a formula imposed on silence produces mostly suspicion.
The design group. Elected members of the screening committee, the finance person, the unit leader, union officers where present, and anyone the committees co-opt. Outside help may advise; it does not draft.
What the group decides. The design of the shared reward is the organisation's own. It is made with the people it would cover, from the organisation's own books, and this book does not describe how a gain is calculated or divided. What the group must also decide, and write down before any vote, are the rules around the design, because these are where Chapter 3 found trust is won or lost:
Kim's survey associated survival with employee approval by vote, union participation and a small bonus group; Bullock and Tubbs found employee involvement in design and employees' favourability towards the plan among the predictors of success.6062 Christine Cooper, Bruno Dyck and Norman Frohlich's 1992 article in Administrative Science Quarterly gave its subject in its title: improving the effectiveness of gainsharing through fairness and participation.69 And Luis Gomez-Mejia, Theresa Welbourne and Robert Wiseman's theoretical model of gainsharing, published in 2000, framed a plan as an arrangement in which employees accept a share of compensation risk.70 Read together, these sources point the same way. People are being asked to accept a measure of risk in exchange for a share of something they help create. They will accept it more readily, and defend it longer, when they wrote the rules about how the risk works.
The vote. Everyone the design would cover votes, by secret ballot, on the design and its rules together. The threshold for adoption is set before the ballot and posted. If the vote fails, the design group asks why, in writing, and either revises or stops. A failed vote is information, not defeat. The answer rule, the measure and the committees continue whatever the result. They were never conditional on it.
Who decides. The people covered decide by vote. The sponsor decides whether the organisation can afford the design the group proposes, and says so before the ballot, not after.
Phase seven · The first review, and the plateau (months sixteen to eighteen, and after)
Haute Lumière · The Gain Was Always Shared · Chapter Eight
The first review. Three to six months after any design takes effect, the screening committee, sponsor and union review it against its written rules. The questions are factual. Were the rules kept? Did any outside event move the measure, and was it adjusted for as agreed? Did the answer rule hold, or did attention move to the money? Did employment hold? Did any role disappear because of an implemented idea, and where did the person go? The review is posted.
The plateau. Robert Kaufman's Improshare firms reported gains that rose to a cumulative seventeen and a half per cent by the third year and then levelled off.65 Every system reaches the point where the obvious ideas have been found. Chapter 3 named this as the moment a plan becomes a fixed cost with no visible return, and is cut in the first bad quarter.
The response is to widen what people can see and change. Kim's survey associated retraining and new-employee training with survival, through their effect on performance.60 In practice this means three things.
New arrivals. Every person who joins the unit is told, in their first week, how ideas are answered, where the measure is posted and how the committees work. They are given the secretary's last three months of records to read. A system that is only understood by those who were present at its founding ages with them.
The last phase is not a phase in time. It is a provision made at the start that matters only on the day it is needed.
Plans end because firms are sold, departments are merged, leaders leave and new ones arrive who did not choose the system. Chapter 4 described the supervisor as the closing valve inside a unit; a new owner or chief executive is the closing valve above it. Santa Anna's book, writing about workplaces that move from hierarchy towards distributed authority, gives the reason a plan is vulnerable at exactly this point:
The transition from traditional hierarchy to distributed authority is a developmental process, not a structural flip71
A new leader who has not lived through the development sees only the structure, and a structure whose reasons are invisible looks like cost. Four provisions make the reasons visible.
A written charter. A document, adopted by the sponsor and the screening committee in phase five, that records why the system exists, what it has done, and what any future leader must do before ending it: at least ninety days' notice, a meeting with the screening committee, and a written answer to the committee's reply. It cannot bind a future owner absolutely. It can make ending the system a deliberate act rather than an oversight.
The record. The secretary's ledger, the sealed expectations and the posted reviews are the evidence a new leader will need. Keep them where they will be found, in plain language, with the before-and-after figures for the measure and for employment.
The union, or the elected committee. Where a union helps run the system, its role does not depend on management's continuity. Kim's association between union participation and survival may partly reflect this.60 Where there is no union, the elected committee members should hold the charter jointly with management.
Ownership. The most durable protection is a structure that changes who holds the firm. The companion book in this lineage on employee ownership, employee ownership, reads the evidence on what happens when people own part of the firm as well as share in a measure, and the companion on patient capital, patient capital, reads how ownership horizons shape what firms keep. A leader who wants a participation system to outlast them should read both.
| Phase | Weeks | What happens | Who decides |
|---|---|---|---|
| 1 · Readiness | 1–4 | Origin Reading; data inventory; Five-Critique Pre-mortem; sponsor's four answers | Sponsor; unit leader |
| 2 · Supervisors first | 5–8 | Union officers met; each supervisor met alone; supervisors' role and measure fixed | Unit leader; union |
| 3 · The answer rule | 9–22 | One rule announced; secretary elected; every idea answered in ten working days; first count at week 14; review at week 22 | Unit leader, on the secretary's record |
| 4 · The measure | 23–40 | Measure chosen; books opened; weekly posting; outside events noted; collection; sealed expectation | Measurement group; finance person |
| 5 · Committees | 41–52 | Elections; outside help builds and leaves; answer rule handed to committees; first annual review | People covered; sponsor sets limit |
| 6 · Design and vote | Months 13–15 | Gate checked; design group; rules written; secret ballot | People covered, by vote; sponsor on affordability |
| 7 · Review and plateau | Months 16–18 and after | First review; training across boundaries; second-order ideas; induction | Screening committee; sponsor; union |
| Change of control | From phase 5 | Charter; record; union or committee holds the charter; ownership considered | Sponsor and screening committee |
Read quickly, eighteen months of procedure can look like caution. It is closer to the opposite. Each phase gives the people covered time to discover that the previous one was safe before the next asks more of them: to write an idea and see it answered, to see a number and believe it, to elect a committee and watch it decide something, and only then to accept a measure of risk in a design they helped write. Phase by phase, the organisation is building the circulation Chapter 4 described, and building it in the order in which trust can carry it.
Chapter 9 turns to the other half of the practice: how to know whether any of it worked, and how to read the evidence honestly when the result is mixed. Chapter 10 gives the conversations this sequence requires, with the board, the union, the supervisors and the sceptics, in words a leader can use. The Appendix gathers every Tool in this book into one workbook, with a ninety-day plan that matches phases one to three.
The leader who runs this sequence will spend most of the first year doing small things: reading cards, writing answers, posting a number, sitting in meetings where someone from the night shift explains a machine. That is the work. The cheque, if one ever comes, will only confirm what the record already shows.
One sheet, kept by the unit leader and the secretary together, reviewed at the end of each phase. Every line is a date, a name or a count.
Roles (named in week one)
Sponsor: · committed until (date, at least two years):
Unit leader:
Supervisors:
Union officers, if any:
Secretary (elected week nine): · protected hours per week: ___
Finance person:
Outside help, if any: · leaving date written in their terms:
Phase one · Readiness (weeks 1–4)
Origin Reading done (date): · with whom:
Operational data available, measure by measure, and months of history:
Pre-mortem done (date): · critiques without a preventing step:
Sponsor's four answers: two-year commitment yes / no · promise on removed jobs (in words): · financially sound yes / no · known inequities and the date they will be fixed:
Decision to begin: yes / not yet · if not yet, the reason and the work for the next ninety days:
Phase two · Supervisors first (weeks 5–8)
Union officers met (date): · what they asked for:
Phase three · The answer rule (weeks 9–22)
Announcement date: · the five points said by shift: yes / no
Answer window: ___ working days · where answers are posted:
Phase four · The measure (weeks 23–40)
Measure chosen: · who counts it: · books opened to all (date):
Outside events that move it, and how each will be noted:
Sealed expectation filed (date): · where:
Phase five · Committees (weeks 41–52)
Elections held (date): · committees formed: ___
Committee budget limit per idea, set by sponsor:
Outside help left on (date): · as planned yes / no
First annual review (date): · expectation opened and compared yes / no · posted yes / no
Phase six · Design and vote (months 13–15)
Design group members:
Sponsor's statement on affordability, before ballot (date):
Ballot date: · threshold set in advance: · result:
Phase seven · Review and plateau (months 16–18)
First review (date): · rules kept yes / no · employment held yes / no · roles removed by ideas and where each person went:
Boundary training started (date): · committee rotation yes / no
Change of control
Charter adopted (date): · held jointly by:
Record stored at:
Ownership options considered (date, and by whom):
Chapter Nine
How to Know It Worked
Haute Lumière · The Gain Was Always Shared · Chapter Nine
Every leader who starts a participation system will, at some point, be asked whether it worked. The question usually arrives late: at a budget meeting eighteen months in, from a new chief executive, from a board member who was not there at the start. By then the honest answer is often unavailable, because nobody wrote down what the organisation looked like before, what it expected, or what else happened in the meantime. The leader is left to answer from impression, and impression is the weakest evidence there is.
This chapter is about answering that question well, and about arranging matters at the start so that it can be answered at all. It is the measurement companion to the field manual in Chapter 8. It covers what to count, how to establish a before, how to read a change against the background of everything else that moves, how researchers have done this in the published studies, and how to read your own results with the same honesty this book has tried to apply to theirs.
Ammanuel Santa Anna's Ecstatic Economics has a passage on national wellbeing budgets that serves as a fair warning for the whole subject. After describing what such budgets measure and why it matters, it turns to their limits:
The framework is necessary but not sufficient.72
A measurement plan is the same. It will not make a participation system work. It will tell you, with some confidence, whether it did, and where it did not. That is worth a great deal, because a system whose effects cannot be shown will be cut by the first person who doubts it.
There are four families of measure, and a participation system should keep at least one from each.
The vital signs of the circulation. These are the measures Chapter 4 named for the up-current and the answer, and the secretary in Chapter 8 keeps them from the first week. Ideas received per month. The number of different people who contributed. The share answered within the agreed time. The share declined with a reason. The share implemented. The share of implemented ideas whose effect was reported back to the person who had them. And, from the work of Jeffrey Arthur and Lynda Aiman-Smith on suggestions over time, the share of ideas that question a process rather than fix a fault within it.73
These are the earliest signals. They move within weeks. If they do not move, nothing else will, and a leader who watches only the outcome measures will learn this months too late.
The work. The measure chosen in phase four of Chapter 8, and its neighbours: defects or errors per thousand units of output, rework, minutes of downtime, waiting time, complaints, output per hour where output can be honestly counted. Robert Kaufman's survey of 112 Improshare firms and Larry Hatcher and Timothy Ross's single-plant study both located much of the gain in quality and in machines standing idle less often.7475 These are likely to move before output per hour does.
The people. Employment, voluntary turnover, absence, grievances, accidents. Michael Schuster's study of up to nine years of a Scanlon plant found employment and turnover stable while productivity rose, and he presented that stability as a finding in its own right.76 Hatcher and Ross reported a significant fall in grievances at the moment their plant changed its pay system.75 And Jeffrey Arthur and Gregory Jelf published a study in the Journal of Labor Research in 1999 whose title states its subject: the effects of gainsharing on grievance rates and absenteeism over time.77 These measures matter for a reason beyond what they show about the system. A productivity gain achieved by cutting people, or by wearing them down, has moved value from one group to another rather than creating it. Counting the people is how an organisation knows which happened.
The relationship. Attitudes, trust, commitment, intention to leave, the sense that it is safe to say a process is wrong. These are measured by survey, and surveys are the weakest of the four families because people answer them strategically and because the questions shape the answers. They are still worth keeping, for one reason: the most striking finding in this literature is about them. Susan Hanlon, David Meyer and Robert Taylor found, three months after a group bonus was withdrawn, higher moral commitment, more prosocial behaviour and less intention to leave among the people who had been under the plan than among a control group.78 Chapter 4 read that as the circulation outliving the money. It could only have been seen because somebody measured attitudes before, during and after.
Santa Anna's book, in the passage on wellbeing budgets, makes a claim about this family that a measurement plan can test rather than assume:
Trust, it turns out, is extraordinarily efficient. Suspicion is extraordinarily expensive.72
The sentence is the author's judgement, not a measured result. A unit that keeps measures from all four families is in a position to see, over two or three years, whether it holds in its own case: whether the months in which the vital signs were strongest were also the months in which grievances fell and defects fell with them.
The single most common error in judging any change inside an organisation is to compare one average after with one average before and call the difference the effect.
The difficulty is that almost everything moves. Defect rates drift with the age of equipment. Output varies by season. Grievances rise in the months after a reorganisation elsewhere. A new supplier, a new product, a new manager two floors up: each can move a number by more than a year of good ideas. Compare one quarter before with one quarter after, and you may be measuring any of these.
The researchers whose work this book relies on dealt with this in two ways, and a leader can borrow both.
Schuster's study of the Scanlon plant, Hatcher and Ross's study of the plant that left piecework, and a 1988 study by John Wagner, Paul Rubin and Thomas Callahan whose title describes its method as an interrupted time series analysis of magnitude and trend, all used the same family of technique.767579
In plain words, an interrupted time series works like this. Collect the measure at regular intervals, monthly or weekly, for a long period before the change. That period shows not only the level of the measure but its trend and its ordinary variation: whether it was already improving, how much it wobbles from month to month, whether it has a seasonal pattern. Then make the change and keep collecting. Now ask two questions. Did the level step up or down at the moment of the change, by more than its ordinary wobble? And did the trend after the change differ from the trend before it?
A step at the moment of the change is harder to explain away than a slow drift, because most background forces do not happen to begin on the day a system is introduced. This is why Schuster's word "abrupt", discussed in Chapter 5, carries weight. The level of output per hour stepped at the point the plan began. And a change in trend tells you something different: that the system did not merely produce a one-off gain but altered the rate at which things improve, which is what a working circulation of ideas should do.
The technique has limits a leader should know. It needs a long before: researchers typically want many data points on each side of the change, which is why Chapter 8 asks for at least three months of collection before the committees are formed and prefers a year. It cannot rule out a background event that happens to coincide with the change. And when several things change at once, as they did in Schuster's plant, where the committees and the bonus began together, it cannot tell you which of them produced the effect.
A unit can do a simple version of this without statistical software. Plot the measure, month by month, for as long a before as exists. Mark the date of each phase in Chapter 8: the answer rule, the posted measure, the committees, any design adopted after a vote. Mark every outside event the measurement group noted. Then look. A line that was flat and steps down after the answer rule, with no outside event at that point, is a different story from a line that was already falling at the same rate before anything began.
The stronger design is a comparison group: a similar unit, at the same time, that did not adopt the system.
Susan Hanlon and Robert Taylor's 1991 study of a self-designed gainsharing plan at a priority package delivery firm, described in its abstract as the first reported gainsharing study to use control groups and pre- and post-tests in a quasi-experimental field study, found that in the first six months positive job-related communication and the communication of ideas increased among the people under the plan.80 The 1994 follow-up with David Meyer is the study, cited above, that kept measuring after the bonus was withdrawn.78 And M. M. Petty, Bart Singleton and David Connell published in the Journal of Applied Psychology in 1992 what their title calls an experimental evaluation of an organisational incentive plan in the electric utility industry.81
A comparison group answers the question an interrupted time series cannot: what would have happened anyway? If both units' defect rates fell by the same amount in the same months, something other than the system was at work. If only the unit with the system improved, the case is stronger.
Most organisations cannot randomise which units adopt a system, and should not try. But many have a natural comparison: a second site, a second shift, a department that will start a year later. Where one exists, collect the same measures in it, from the same date, with the same care. Starting the system in one unit first and another a year later is not only a sensible way to learn; it gives the organisation its own comparison for free.
A comparison is never perfect. The comparison unit knows it is being compared, which changes behaviour. Managers talk to each other, and good practices spread. Units differ in ways nobody listed. The honest reading of a comparison is that it narrows the range of explanations. It does not close it.
The survivors, again
Haute Lumière · The Gain Was Always Shared · Chapter Nine
Chapter 3 named the fifth critique of gainsharing as a problem in the evidence itself: case studies are published more often when the plan worked, and surveys of current users cannot reach the firms that gave up. Kaufman's 112 Improshare firms were those still known to be using the plan in 1988 who chose to reply.74 Bullock and Tubbs's thirty-three cases were the ones somebody published.82
The same bias operates inside an organisation, and it is worth naming because it is the one most leaders do not see.
Inside a unit, the ideas that get remembered are the ones that worked. The committee meetings that get described are the good ones. The supervisor who resisted and was moved is no longer in the room when the system is reviewed. The people who left in the first year, perhaps because the system asked something of them they did not want to give, perhaps because it did not ask enough, are not surveyed in the second. Every review of a system conducted only from the inside, by the people who remain, is a survey of survivors.
Three habits counter it.
Keep the declined ideas. The secretary's ledger in Chapter 8 records every idea and every answer, not only the adopted ones. A review that reads the declined ideas, and asks how many were declined for good reasons and how many because nobody wanted to deal with them, sees what a review of successes cannot.
Ask the leavers. Every person who leaves the unit in the first two years is asked, in a short conversation with someone outside their line, what they thought of the system and whether it had anything to do with their leaving. Most will say no. The ones who say yes are the evidence a survivor survey cannot reach.
Count the stops. Committees that stopped meeting. Supervisors who stopped answering. Months in which the measure was not posted. Every interruption is recorded with its date and reason. A system that has never been interrupted is either very well run or not being watched closely.
When Kaufman reported Improshare's results, he noted that mean productivity increases were larger than median increases and followed the same pattern.74 Chapter 5 drew the practical lesson: if you use one figure from that study to think about your own firm, use the median, and treat even that as an upper estimate.
The distinction matters inside a unit too. Suppose a plant has six lines. Five improve a little. One improves a great deal, because its machine had a fault that a single idea fixed. The average improvement across six lines looks impressive. The typical line improved a little. Both statements are true. A leader who reports only the average will be believed until someone looks at the lines one by one, and then will not be believed about anything.
Report, at minimum:
That last line is the habit of every note in this book, and a leader can adopt it without any training in statistics. A figure accompanied by a plain statement of its limits is more persuasive to a sceptical board than a larger figure without one, because it shows the person presenting it has already asked the board's question.
Writing the expectation down first
Haute Lumière · The Gain Was Always Shared · Chapter Nine
Chapter 8 asked the unit leader, secretary and finance person to write down, before the committees were formed, what they expected to change over the next year, by how much and by when, and to file that page where it could not be revised. It is worth saying why this single page matters so much.
Without it, every result can be read as a success. If defects fell by a tenth, that was the goal. If they fell by a fiftieth, the real goal was cultural. If they did not fall at all, the system has laid foundations whose effects will show later. Each of these readings may be true. None can be checked, because the expectation was formed after the result was known.
With it, the review has a fixed point. Expected ideas per month against actual. Expected median days to answer against actual. Expected change in the measure against actual. Where actual fell short, the review asks why. Where it exceeded expectation, the review asks whether something else was happening. Either way, the organisation learns something it could not have learned by looking only at the result.
This is the discipline of a pre-registered study, applied to a single unit. It costs one page and one hour. It is the cheapest protection available against the most common error in judging change, which is deciding what you meant after you see what happened.
Most results will be mixed. The vital signs will rise and the measure will not move. Or the measure will improve and grievances will rise. Or everything will improve in one unit and nothing in another.
Chapter 4 offered a reading that helps here: a failing system is not a verdict, it is a location. Read a mixed result by asking where the circulation has blocked.
Santa Anna's book offers a personal practice that maps onto this reading at the scale of one life. In the weekly Vitality Inventory, the reader reviews the economic events of the week and asks of each whether it generated or depleted vitality, and the book says the practice builds a map more reliable than a spreadsheet because it measures
the quality of aliveness that your economic life produces.83
The claim about reliability is the author's. The structure of the practice is useful for a unit. A quarterly review that asks of each measure not only whether it rose but whether the people producing it are more or less able to keep producing it is asking the question the people measures exist to answer. Chapter 11 returns to the Inventory at the scale of a household.
The moment this chapter began with, the question from a new chief executive or a sceptical board, deserves a final word, because it is where measurement either protects a system or fails to. Chapter 10 gives the conversation in full; here is what to bring to it.
Bring the line, not the average: the monthly measure plotted from the earliest data, with each phase and each outside event marked. Bring the vital signs, month by month. Bring the people measures. Bring the sealed expectation and the comparison with what happened. Bring the comparison unit, if there is one. Bring the declined ideas and the leavers' answers. And bring, in a single paragraph, what none of this can show.
A system presented this way does not need to claim more than it achieved. It needs only to show that the organisation has been watching itself honestly. That is rare enough to be persuasive in its own right.
A quarterly review, two hours, conducted by the screening committee with the sponsor. Fill every line from the secretary's ledger and the posted measure. Where a figure is an estimate, mark it as one.
1 · The before
Was the measure already improving before anything began? yes / no / unclear · evidence:
Sealed expectation opened (date filed):
2 · The vital signs this quarter
3 · The work
Did the line step at any phase? which: · by more than its ordinary month-to-month variation? yes / no
Did the trend change after any phase?
4 · The people
Roles removed by implemented ideas ___ · where each person went:
5 · The comparison
Comparison unit, if any: · same measures collected from the same date? yes / no
Change in the comparison unit over the same period:
6 · The survivors
Leavers asked about the system: ___ of ___ · said it played a part: ___ · what they said:
Interruptions this quarter (committee missed, measure not posted, answers late), with dates:
7 · Expected against actual
| Measure | Expected (sealed) | Actual | Gap | Why |
|---|---|---|---|---|
| Ideas per month | ||||
| Median days to answer | ||||
| Share implemented | ||||
| The chosen measure | ||||
| Voluntary turnover |
8 · Where the circulation blocks
One change for next quarter: · owner:
9 · What this reading cannot show
In one paragraph, the alternative explanations this quarter's evidence cannot rule out:
Chapter Ten
The Hard Conversations
Haute Lumière · The Gain Was Always Shared · Chapter Ten
Every participation system is first a series of conversations, and most of them are harder than the design work. A board has to hear that the firm intends to open part of its books. A union has to hear that management wants its help with productivity, a word with a long and unhappy history. A supervisor has to hear that ideas will now travel past their desk. A partner in a small firm has to hear that the other partner wants to share something both of them have always kept. A family that owns a business has to hear that the people who are not family will see numbers the family has never shown them. And somewhere in the room there is always a sceptic, usually well read, who has a good argument for why none of this will work.
This chapter is about those conversations. It does not offer techniques for winning them. A participation system that is sold by persuasion and accepted under pressure has already failed the test Chapter 2 set out, because it began by asking nobody anything. What follows is a way of entering each conversation honestly: what the other person is most likely to be afraid of, what the evidence can and cannot say to that fear, and a short script, written as plain spoken lines, that a leader can adapt in their own words.
Ammanuel Santa Anna's Ecstatic Economics is frank about the difficulty, in a passage about cooperatives and steward ownership that applies here without change:
Anyone who tells you otherwise is selling something.84
The scripts below are not meant to make the conversations easy. They are meant to make them true.
Each script is short on purpose. It names what the speaker wants, what they do not know, and what they are asking the other person to decide. None of them contains a figure that the firm has not measured, a promise about results, or an appeal to urgency. Where the scripts mention research, they mention it at the strength Chapters 3 and 5 established: associations, single sites, survivors. A leader who overstates the evidence in the first conversation will spend the next year being held to the overstatement.
The people in the examples are composites, built for teaching from the kinds of situations the research describes. None is a real person and none of the lines attributed to them are quotations.
What they fear. Three things, usually in this order. That opening any part of the accounts will invite claims the firm cannot meet. That a shared reward will become a fixed cost that cannot be withdrawn in a bad year. And that management is about to spend time and money on something soft.
What the evidence can say. It can say that the most careful case review in the literature, Bullock and Tubbs's analysis of thirty-three published cases, found success robust across firm size, union status, technology and environment, and found that the features predicting it were mostly about participation in design.85 It can say that Dong-One Kim's survey of 211 organisations associated programme survival with, among other things, sound financial health and major capital investment alongside the plan.86 And it can say, honestly, that both sources over-represent firms where things went well. It cannot say what will happen in this firm.
What the board actually needs to decide. Not a reward. Chapter 8 places the design of any shared reward at the end of the first year, after the answer rule, the visible measure and the committees have been tested. The decision in front of the board at the start is much smaller: whether to let management run the first half of Scanlon's idea, the half that needs no change in pay, and to review it at a fixed date.
The script.
We are not asking you to approve a bonus scheme today. We are asking for twelve months to do three things that cost very little: answer every idea from the floor in writing within ten days, put one measure of quality where the people who affect it can see it, and form committees to read the ideas. We will report to you each quarter on ideas received, answered and implemented, and on the measure. At the end of the year we will come back with what we found. If the case for a shared reward is there, we will bring a design that the people it covers have helped to write, with written rules for how it can change. If the case is not there, we will say so and keep what worked. The research says plans like this succeed most often when the people covered help design them, and survive more often when they vote on them. It also says the studies mostly see the firms that succeeded. We are not promising you a result. We are proposing to find out, carefully, with a record you can read.
The last sentence matters most. A board that has been promised a result will treat the first flat quarter as a failure. A board that has been promised a record will read it.
When an owner says no. Ask what would have to be true for a yes, and write the answer down. Then do the parts that need no board approval: the answer rule within your own unit, and the visible measure. Chapter 4 argued that the answer is the most frequent payment a firm makes to its people and the least expensive. It does not need a resolution.
What they fear. That gainsharing is a way to buy concessions: wages held down in exchange for a bonus that may never pay. That committees will be used to bypass the bargaining relationship. That ideas offered in good faith will be used to remove jobs. Each of these has happened often enough to be a reasonable fear.
Kenneth Thornicroft's ethnographic study of one American firm's Scanlon plan, published in 1991, argued that such plans should not be treated as simple trades for collective bargaining concessions, because they necessarily disturb the status quo, and that success needs both a sound structure and a change in culture through which all employees can take part meaningfully in decisions.87 A union officer who has read that study, or lived something like it, is right to ask which kind of plan is being offered.
What the evidence can say. Kim's survey associated union participation in running the programme with survival.86 Jeffrey Arthur and Dong-One Kim's comparison of two locations with similarly structured plans proposed that employees' willingness to share high-risk knowledge with management depends on the degree of labour–management cooperation and trust, and that union support and involvement increases employees' acceptance of the risk that a variable reward carries; their results, they report, generally supported the hypothesis.88 Two sites is a small base. The direction, though, is consistent with the origin story in Chapter 2: the first plan was a union officer's proposal.
What the union needs to hear. That the plan will not replace bargaining, that the union will be a party to its design and its running, and what will happen to a job that an idea removes. Chapter 5 noted that the plant in Michael Schuster's study of up to nine years kept employment and turnover stable while productivity rose.89 That is one plant, and it is not a guarantee. But a firm that cannot make some version of that commitment should not expect the union, or anyone, to send it the ideas that matter.
The script.
This is not a substitute for the contract and we will not ask you to trade anything in it for this. We want the union in the design from the first meeting, with seats on every committee and a say in the rules for how any measure is set and changed. Before anything else, we will put in writing what happens to someone whose work is made unnecessary by an idea, and we want you to help write it. We will not design any shared reward for a year. In that year we will answer every idea in writing and show the measure every week. If at the end of it you do not think the answers were real, tell us, and we will not go further. The research we have read suggests plans like this last longer when the union helps run them. We would rather build it with you than present it to you.
When the union says not yet. Take it at its word. Ask what it would need to see in six months, and show it. A union that watches a firm answer every idea for six months, including the awkward ones, has better evidence than any study.
The supervisors
Haute Lumière · The Gain Was Always Shared · Chapter Ten
Chapter 3 named the supervisor as the person a participation system most often threatens, and the reasons are structural. A supervisor's authority usually rests on being the channel through which problems are reported and decisions are made. A working committee opens a second channel. Some ideas will describe problems the supervisor knew about and did not fix.
What they fear. Exposure, and irrelevance. Both are reasonable.
What the book says about it. Santa Anna's Ecstatic Economics, writing about firms that move from hierarchy towards distributed authority, puts the difficulty plainly:
You cannot distribute authority to nervous systems that are still organized around fear.90
A supervisor who expects to be blamed will defend, quietly and effectively. The conversation has to come first, one supervisor at a time, before anyone else hears the plan.
The script.
Before we tell anyone else, I want to talk to you, because this changes your job more than anyone's. Ideas from your team are going to start coming in on cards and through a committee. Some of them will be about things you already knew were wrong and could not get fixed. I want you to be the person who brings those ideas forward, and I want your name on the board next to your team's names when one is used. For the first ninety days you will be measured on one thing: how many of your team's ideas get an answer on time. Not on how many problems they find. What would you need from me for that to be true? What are you worried about?
Then listen, and write down the answer. Supervisors usually ask for three things: time to attend committees, authority to fix small problems without waiting, and protection from being blamed for problems the cards reveal. All three are reasonable, and the third is the one that decides whether the plan survives.
The person who keeps the accounts is often the last to be consulted and the first to see the risk. Opening books, even partly, means choosing what to show, how often, and at what level of detail, and it means that every number shown will be questioned.
What they fear. Misreading. A number shown without context will be read as evidence of something it does not show. And exposure: margins, pay, the cost of a decision made at the top.
What the book says. Santa Anna, writing about pay transparency, is direct about the risk of showing an unfair system:
you can't make an unjust system transparent and expect positive results.91
and, in the same discussion, about why visibility matters at all:
The nervous system relaxes when it can see the landscape clearly.92
Both are true at once. A finance lead who hears only the second sentence will open too much too soon. One who hears only the first will open nothing.
What to show first. Not the accounts. Chapter 8 starts with one operational measure that the group can affect: defects, downtime, rework, waiting time. That number usually comes from operations rather than finance, which makes the first step easier. The broader financial picture can follow, as a monthly briefing that explains what moved and why, prepared with finance and given by finance.
The script.
I am not asking you to open the ledger. For the first year, the only number we will post is defects per thousand, by line, weekly. What I am asking is for your help in deciding what we will show after that, and how. You know better than I do which numbers will be misread. I would like you to write down, now, what you would be comfortable showing in twelve months, what you would need to explain it properly, and anything you think we should fix before anyone sees it. If there is something in the pay structure that would look unfair under daylight, I would rather we found it together than had it found for us.
In a practice of two or five, the conversation is closer and more personal. Partners have usually kept the firm's results among themselves. Sharing any part of a gain with employees, or even showing them a measure, feels like giving away something that was hard to earn.
What they fear. Loss of control. That the people who work for them will start to feel entitled to decisions. That the firm will become slower.
What can honestly be said. That the evidence in this book is mostly from plants of a hundred or more people, and that nobody has studied a practice like theirs in the same way. Kaufman's Improshare survey did find that smaller firms gained more than larger ones, but that is a finding about self-reported results among surviving users, not a forecast.93 What can be said more confidently is that the first half of the idea, asking and answering, costs a small firm almost nothing and can be stopped at any time.
The script.
I want to try something for three months and I want us to agree on it first. Every idea anyone here has about how we work gets an answer from one of us, in writing, within a week. We pick one number we both care about, say how many jobs go back for correction, and we put it on the wall. At the end of three months we look at it together and decide whether to keep going. Nothing about pay changes. If you think it is slowing us down, say so at the end and we stop. I am asking because I think the people here know things about our work that neither of us sees, and I want to find out whether that is true.
The family business
Haute Lumière · The Gain Was Always Shared · Chapter Ten
Here the hard conversation is often among the family before it reaches anyone else. A family that owns a firm carries its own history of money, which Santa Anna's book treats as something the body remembers. One of the book's reflection questions asks:
What economic somatic patterns did you inherit from your family of origin?94
In a family firm, those patterns run the business. A founder who grew up with the memory of nearly losing everything may experience any opening of the books as a return to that danger. A second generation may want to share more and feel unable to say so to a parent. The conversation is less about gainsharing than about who the firm is for.
What they fear. That family wealth will be diluted, that private matters will become public among employees, and that the next generation will lose the firm.
What can be said. That the first year described in Chapter 8 does not touch ownership or family wealth at all. That a written charter for a participation system, decided now, can protect it through a change of generation in a way an informal habit cannot; Chapters 4 and 8 noted that plans end when firms are sold or control changes. And that the neighbouring book in this series on employee ownership, employee ownership, reads what happens when families decide to go further.
The script, for the family meeting.
I want us to talk about something before we talk to anyone outside this room. I would like the people who work here to be able to send us their ideas and get an answer, and to see one number about our quality every week. I am not proposing to share ownership or to show anyone our private accounts. I know this feels different to each of us, and I think some of that comes from how each of us grew up with money in this family. I would like each of us to say what worries us about it. Then I would like us to decide whether to try it for a year, and if we do, to write down how it would continue if one of us was no longer running the business.
Every room has one, and they are useful. The sceptic's arguments are usually good, and a leader who answers them honestly will have a better plan than one who never heard them.
The free-rider argument. Eugene Kandel and Edward Lazear set it out in 1992: in an organisation of any size, a person who works harder for a shared reward receives only a small fraction of the value their effort creates, and free-rider effects would seem to choke off the motivational force of any group incentive.95 The honest reply begins by agreeing. Then it notes that their article goes on to analyse how peer pressure, norms and mutual monitoring change that arithmetic, and that the effect depends on whether people can see each other's work. Theresa Welbourne and Claudia Ferrante's study of 203 employees in a customs brokerage and freight-forwarding firm found that what they called advisory monitoring, reacting to co-workers' behaviour, was positively related to managers' ratings of workers' performance under gainsharing and negatively related under traditional merit pay alone.96 One firm, and ratings rather than output. But it suggests that the pay system changes how people treat each other's attention.
The piece-rate argument. Lazear's study of Safelite Glass found that output per worker rose by about forty-four per cent when installers moved from hourly pay to piece rates.97 The honest reply agrees that piece rates suit work done alone and counted exactly, as fitting a windshield is. It then asks the sceptic which of the firm's problems are like that. Much of the gain in Chapter 5's Improshare firms came from fewer defects and less downtime, which are rarely the product of one person.
The survivorship argument. The sceptic who has read Chapter 3 will say that the evidence sees the survivors. They are right. The reply is Chapter 9: treat the firm's own plan as the study, collect the measure before anything is shared, and keep the record of what fails.
The script.
You are right about the free-rider problem, and the people who described it best say the same thing you do. They also say it changes when people can see each other's work, so that is what we are going to build first. You are right that piece rates work for some jobs. Most of what goes wrong here is not one person working slowly; it is defects and machines standing idle, and those take more than one person to fix. And you are right that the published research mostly sees the plans that worked. That is why we are going to collect our own numbers for a year before we share anything, and why I would like you on the committee that reads them. If it is not working, I want the person most likely to notice to be in the room.
The last invitation is the most important line in the chapter. A sceptic given a seat becomes the plan's most useful instrument. A sceptic left outside becomes its opposition.
Read together, the seven conversations follow one pattern. Each begins by asking for less than the listener feared. Each names what is not known. Each puts a decision in the listener's hands and a date on it. And each ends by inviting the listener into the work rather than asking them to approve it.
That pattern is not a sales method. It is the participation system in miniature: the question asked first, the answer owed, the measure made visible, the people affected included in the rule. A leader who cannot hold these conversations in this way is unlikely to run a system that holds them a thousand times a year on cards and at committee tables.
Santa Anna's book, discussing what governments signal when they choose what to measure, ends a paragraph with two short sentences that could stand over each of these meetings:
Trust, it turns out, is extraordinarily efficient. Suspicion is extraordinarily expensive.98
The sentence is the author's judgement, not a measured result. In a hard conversation it is a useful reminder of what is being built, one honest meeting at a time.
The fuller version of these conversations, and the books on ownership and reciprocity that sit beside this one, are part of the house's library at hautelumiere.life; the lens behind all of them is Santa Anna's Ecstatic Economics, at Ecstatic Economics. The gift and reciprocity book in this series, gift and reciprocity, reads the experiments that help explain why an invitation is answered more often than a demand.
Use one sheet for each hard conversation, filled before the meeting and completed after it. Keep the sheets together; they are part of the record Chapter 9 asks a firm to keep.
Before
Who: · role: board / owner / union / supervisor / finance / partner / family / sceptic
What this person is most likely to fear, in their terms, not ours:
What the evidence can honestly say to that fear, and its limit:
What we are asking them to decide, in one sentence:
What we are not asking for (write it, so it can be said aloud):
The date by which the decision would be reviewed:
Any figure we plan to mention, and where it was measured: (if none was measured, mention none)
The script, in our own words (five to eight sentences)
What we want:
What we do not know:
What we will show them, and how often:
What happens to a job an idea removes (if relevant):
The invitation into the work:
During
Their first concern, in their words:
What they would need for a yes:
Anything they told us that changes the plan:
After
Decision: yes / not yet / no · conditions:
What we committed to, and by when: · owner:
Next conversation with this person:
If not yet or no: the part of the plan we can still do without their approval:
Chapter Eleven
One Person's Gain
Haute Lumière · The Gain Was Always Shared · Chapter Eleven
Most of this book has been about plants of a hundred or more people, because that is where the research was done. But the idea at its centre does not depend on size. A firm that asks the people closest to the work what they know, answers them, makes one number visible and returns some of what is saved is doing something a household can do at a kitchen table and a single person can do with their own working week.
This chapter brings the idea down to that scale. It is written for a reader who does not run a plant: someone who shares a household budget, works as an employee inside a firm they do not control, or works alone. It offers no advice about investments or savings products, and it promises no particular result. What it offers is a way of reading one's own economic life through the same four questions the earlier chapters asked of firms: who knows what, who answers, what can be seen, and what comes back.
There is a reason to do this beyond convenience. Ammanuel Santa Anna's Ecstatic Economics argues that economic life is not only something that happens in institutions but something carried in each person. In the book's chapter on the somatic dimension of economics, it puts that claim in a sentence:
The economy you live in is not just a system of policies, institutions, and market mechanisms. It is a pattern in your body.99
A leader who has never examined how they themselves react when money is discussed will find it hard to build a firm that answers its people calmly. The household and the person are where the practice is learned.
A household that shares money is, in the terms Chapter 12 uses, a small commons. Several people draw on a shared resource. Several people can make it last longer or run it down. Nobody can manage it well alone, and the benefit of care is shared whether or not the care is.
Elinor Ostrom's design principles, read in this lineage's book on the commons, The Pasture Was Never Empty, translate almost directly. The boundary is clear: the people who share the budget. The rules work best when they fit local conditions, which in a household means the particular habits and pressures of these people rather than a template from a personal finance column. The people affected take part in making the rules. Monitoring is done by the users themselves. Small breaches are met with small responses. And there is a cheap, regular way of talking about disagreements before they become large.
Most households do some of this and not the rest. Commonly, one person holds the information and makes the decisions, and the others learn the result when something has gone wrong. That is the household version of the firm Chapter 2 described before Scanlon: the people closest to the costs are never asked, and their knowledge stays where it is.
Who knows what. In most households the person who pays the bills is not the person who notices the costs. A partner sees the heating left on in an empty room. A teenager knows which subscription nobody opens. An older relative remembers which repair would have been cheaper if it had been done a year earlier. A hypothetical household, unnamed and built only for illustration, might discover that the person most aware of food waste is the one who cooks least often, because they are the one who clears the fridge.
The first step is the one Scanlon took in the mill. Ask. Not in a crisis and not as an accusation, but as a regular question: what do you notice about where our money goes that the rest of us might not?
Who answers. An idea offered at home and met with silence teaches the same lesson it teaches on a factory floor. The second time, it is not offered. The household version of the answer rule is simple: every suggestion gets a reply within a week, even if the reply is "not now, because". It costs a few minutes. It is also, as Chapter 4 argued about firms, the most frequent payment a group makes to its members, and the one most often neglected.
What can be seen. A household needs one number that everyone can see and that everyone can affect. Not the whole budget, which can feel like an exam, and not the bank balance, which moves for too many reasons. Something narrower: the monthly energy bill, the food bill, the amount set aside at the end of the month. Chapter 3 warned that a number which moves for reasons nobody controls teaches people that the number is weather. A household number should be chosen so that the people living there can see their own effect on it.
What comes back. When the number improves, something returns to the people who improved it. The book is careful not to prescribe what. In some households it is a shared outing, in some it is a contribution to something one member cares about, in some it is simply a sentence of recognition at dinner naming who noticed what. The point, as in Chapter 4, is that the return is felt by individuals even when the gain is shared. A household that says "the bill came down" has returned a measure. A household that says "the bill came down after you noticed the thermostat schedule" has returned an answer, a measure and a recognition at once.
There is a reason these conversations can be harder at home than at work. Santa Anna's book describes the body as holding a record of every economic experience a person has had, and lists some of the earliest entries:
The first time you watched your parents argue about bills.100
Two adults sharing a household bring two such records. One may have grown up where money was never discussed and any mention of it signalled trouble. Another may have grown up where it was discussed constantly and anxiously. When they sit down to choose a household number, each is reacting partly to the present and partly to a room they left decades ago.
The book's reflection questions make this explicit. One asks:
What economic somatic patterns did you inherit from your family of origin?101
The question continues by asking not only about what was said about money but about how bodies changed when it was discussed: posture, the energy in the room, how parents looked when they worked. A household that wants to run its money as a small commons can begin there. Before any number is chosen, each person can say, briefly, what money felt like in the house where they grew up. It is not therapy and should not be treated as one. It is a way of making visible the thing that will otherwise shape every later conversation without being named.
This is the household equivalent of Chapter 10's family business conversation. The difference is only scale.
Reading your own workplace
Haute Lumière · The Gain Was Always Shared · Chapter Eleven
Most readers of this book are not in a position to install a participation system. They work inside a firm someone else runs. The four questions are still useful, because they let a person read the circulation they are part of and decide where to put their attention.
Take the Circulation Map from Chapter 4 and fill it from your own position.
The up-current. Is there a route by which your ideas about cost, quality or waste can reach someone who can act on them? Is it formal or informal? When did you last use it?
The answer. When you last offered an idea, did anything come back? How long did it take? Was a refusal explained?
The measure. Is there a number you can see that tells you whether your work, or your team's, made a difference? Who shows it to you?
The share. When your team improved something, what came back: recognition, time, training, money, nothing? Has a target ever been raised because it was beaten?
The answers will not change the firm. They will tell you something about it. A person who finds that their ideas travel and are answered, even without any shared reward, is working in a firm whose circulation is partly open. A person who finds that ideas vanish, nothing is visible and targets ratchet is working in a firm where the current runs one way. Neither finding is a verdict on the person. Both are information about where effort is likely to be met.
Sometimes the reading suggests a request. An employee cannot install an answer rule, but they can ask for one within their own team, and the request is easier to grant than most.
A hypothetical employee, unnamed, working on a small team, might put it to a manager like this:
I have had a few ideas about how we could cut rework on our team. I would like to send them to you in writing and get a short answer back, even if it is no. Could we try that for a month? I will keep a note of what I send and what happens, and we can look at it together at the end.
The request is small, specific and reversible. It asks for nothing that costs money. It offers a record. And it changes, in a modest way, the circulation the person works inside. If the manager says yes, the employee has built a channel. If the manager says no, the employee has learned something important about where they work, at very little cost.
Santa Anna's book, writing about leaders who open meetings by naming what they are struggling with, describes the practice as costing about ninety seconds. The same proportion holds for a request like this. It takes very little time to ask, and the answer, either way, is worth having.
A freelancer or sole trader is a firm of one, and the circulation runs between the person and their own work. The four questions still apply, in a slightly altered form.
Who knows what. The person doing the work knows things the person running the business does not, even when they are the same person. The tired version at the end of the week notices which client relationships cost more than they pay, which tasks always take longer than quoted, which tool wastes an hour a day. The planning version on Monday morning rarely asks.
Who answers. A simple practice: keep a short weekly note of what the working self noticed, and have the planning self answer it in writing. Yes, no, not yet. It sounds artificial. It is the same discipline as the answer rule, applied to one person who plays two roles.
What can be seen. One number, chosen because the work can affect it: hours spent on unpaid revisions, days between finishing a job and being paid, the share of the week spent on the work one is best at.
What comes back. When the number improves, return some of the saved value to the person who saved it, which is to say, to oneself, deliberately. Time, rest, training, a piece of equipment that makes the next week easier. Without a return, a person who works alone tends to treat every saving as capacity for more work, which is the household version of the ratchet.
The Vitality Inventory
Haute Lumière · The Gain Was Always Shared · Chapter Eleven
Santa Anna's book offers a weekly practice that suits this chapter closely. It calls it the Vitality Inventory, and it asks a person to review the economic events of the past seven days, purchases, work projects, financial decisions, conversations about money, and to notice for each one their state during it, whether it left them more or less alive, and what their body's honest assessment was. The book describes what the practice produces over time:
Over weeks and months, this inventory builds a somatic map of your economic life102
and names what that map measures:
the quality of aliveness that your economic life produces.103
The book's claim that such a map is more reliable than a financial spreadsheet is the author's judgement, not a measured finding, and this chapter does not make it. What the inventory can do, used beside the four questions, is add a dimension the numbers leave out. A household number that improves while everyone feels more anxious is telling only half the story. A working week that shows a better figure on paid hours while leaving the person depleted has moved one number at the cost of another that matters more.
Used together, the four questions and the inventory give a person a small, complete instrument: what is known, what is answered, what is seen, what returns, and how it all feels.
Chapter 3 read five ways a participation system fails in a firm. Each has a household form, and each has the same way through at a smaller scale.
One share among many. In a household of five, a person who saves on the energy bill sees only part of the benefit, and a person who wastes bears only part of the cost. The free-rider problem is real at home, as anyone who has shared a flat knows. The way through is the one Kandel and Lazear pointed to for firms: people who can see each other's effect on the shared number, and who care what the others think, behave differently from strangers.105 A household number posted where everyone sees it, with a name written beside the change that moved it, does most of the work.
The ratchet. A household that cuts its food bill and is immediately told to cut it further has rebuilt the problem Chapter 3 described. The people who made the saving learn that the reward for effort is a harder target. Agree in advance, together, when the household target will be revisited, and keep to it.
Power. In most households one person holds more of the money and more of the decisions, for reasons that range from income to habit to history. A household that asks for everyone's ideas and then lets one person decline them without explanation has the form of participation without its substance. The answer rule is the test here as it is in a plant: if suggestions stop receiving replies, the system has already closed.
Weather. A household number can move for reasons nobody inside controls: a price rise, a cold winter, an illness. When it does, say so aloud, and do not read the movement as anyone's failure. The choice of number in the Tool below asks for this in advance.
What the record sees. A household that only remembers its successes will learn very little from itself. Keep the page, including the months when nothing improved and the suggestions that turned out to be wrong. After a year it is the household's own study, and it will say more about this household than any general advice can.
None of this requires a household to become an institution. It requires only that the people who share a resource can see it, shape the rules for it, be answered when they speak about it, and trust what they are shown. Those are the same four conditions Chapter 3 found beneath every critique of gainsharing, and they hold at the smallest scale as surely as at the largest.
The household and the person are not a retreat from the larger argument of this book. They are where it is practised first. A leader who has learned to ask and answer at their own kitchen table, to choose a number the household can affect, and to return some of what is saved to the people who saved it, enters the boardroom conversation in Chapter 10 with something better than a theory.
Santa Anna's book makes the larger connection explicit. Describing how gift economies understood wealth, it gives the principle in a sentence this chapter has tried to apply at the smallest scale:
Wealth was not something you had. It was something you moved.104
A household where ideas move, answers come back and saved value returns to the people who saved it is a small place where that sentence is true. It does not need a plan, a committee or a formula. It needs a question asked at dinner, an answer given within the week, and one number on the refrigerator door.
The neighbouring books in this series take the same thought in other directions: gift and reciprocity reads the research on reciprocity that explains why an answered suggestion is worth more than its cost, and Enough by Design reads the economics of enough. The lens behind all of them is Ecstatic Economics, in its colour art edition at Ecstatic Economics, and the house's library is at hautelumiere.life.
One page, for a household or a person working alone. Fill it on the first Sunday of a month and again at the end of the third month. It involves no financial products and no change to anyone's income.
1 · The people
Who shares this budget or this work:
For each person, one sentence on what money felt like in the house they grew up in (optional, said aloud rather than written if preferred):
2 · Who knows what
Three things someone here notices about where money or time goes that the person deciding may not:
When each was last mentioned, and what happened:
3 · The answer
Our rule: every suggestion gets a reply within ___ days, even if it is no. Who replies:
4 · The one number
The number: · today: ___ · where it will be posted:
Things we control that move it: · things we do not:
5 · What comes back
If the number improves, what returns, and to whom:
Has the household or the person ever responded to a saving by expecting more? yes / no · when:
6 · The weekly inventory (one line a week)
| Week | Event | State during it (settled / anxious / shut down) | More or less alive afterwards |
|---|---|---|---|
| 1 | |||
| 2 | |||
| … | |||
| 13 |
7 · At the end of three months
The number, then and now:
What the inventory showed that the number did not:
One change for the next three months:
Chapter Twelve
The Gain Was Always Shared
Haute Lumière · The Gain Was Always Shared · Chapter Twelve
Every improvement in a working organisation has already been paid for once, before anyone designs a reward for it. It was paid for by the person who noticed the sealing jaw drifting after four hours, the nurse who saw that the handover form lost the one line that mattered, the clerk who realised that a report three departments produced was read by none of them. They paid in attention, usually unasked, often on their own time, and in most firms the payment was never recorded.
The question gainsharing asks is not whether to share the gain. The gain is already shared, in the sense that it was produced by many people. The question is whether the firm will acknowledge that, in money and in structure, or whether it will continue to book the result as management's and the knowledge as free.
This book has read the idea from many directions. Its origin in a failing mill and a union officer's question. Its five characteristic failures, each with a way through. The house's reading of it as a circulation, with knowledge moving up and answers moving back. The published cases, every figure traced and graded; the service firm where the first return was talk; the harder rooms where plans broke their promises or found their savings somewhere other than the floor. Then the practice: the field manual, measurement, the hard conversations, the household. This last chapter puts gainsharing beside its neighbours, says plainly what it cannot do alone, gives the living-systems reading of the whole, and ends with what a leader can start on Monday.
Gainsharing is one of several answers the Ecstatic Economics series reads to a single underlying question: when many people produce something together, how do they hold it? Three companion books take up other answers, and each one sharpens what gainsharing is.
Ownership. The companion book on employee ownership and profit sharing reads the research of Douglas Kruse, Joseph Blasi and Richard Freeman, gathered in Shared Capitalism at Work, on firms where employees own shares or share in profit.106 Ownership is a deeper answer than gainsharing in one respect and a shallower one in another. It gives people a claim on the whole firm, including its future, which a gainsharing plan does not. But a share of profit is affected by everything the firm does, including decisions a worker on the night shift cannot see or influence. Gainsharing's narrower measure is, in that sense, a closer one: it ties the reward to what the group can actually change. A firm can have both; the companion book reads the evidence on combinations, and this chapter does not anticipate its conclusions.
Ammanuel Santa Anna's Ecstatic Economics describes the difference ownership makes in terms of what a person is watching for:
The wage employee's neuroception is scanning for threats from the organization — layoffs, pay cuts, arbitrary management decisions.107
The sentence continues by saying that the worker-owner's attention turns towards threats to the organisation instead. Gainsharing, at its best, produces a partial version of that turn. The person who knows the sealing jaw drifts begins to treat the drift as their problem.
Reciprocity. The companion book on gift, reciprocity and trust reads George Akerlof's 1982 account of labour contracts as partial gift exchange and the 1993 experiments of Ernst Fehr and his colleagues on fairness and reciprocity,108 in which people answer generosity with effort and unfairness with withdrawal, even at cost to themselves. That research is the down-current of Chapter 4 studied under laboratory conditions. It suggests why the answer to an idea matters so much: an answer is a small gift of attention, and people reciprocate. It also suggests why a ratchet is so damaging: a raised target after a good year reads as an unfair return, and people withdraw.
Santa Anna's book states the economic consequence of ignoring this in one sentence about markets that treat relationships as costs:
An economy that destroys relationships in order to produce commodities is destroying wealth in order to create the illusion of wealth.109
A gainsharing plan that pays out while the suggestion box goes unanswered is a small instance of that illusion. The bonus looks like sharing. The relationship that produces the knowledge is being spent.
The commons. The companion book on Elinor Ostrom's work reads the design principles of her 1990 Governing the Commons110 for groups that govern a shared resource without either a central owner or a market: clear boundaries, rules that fit local conditions, the people affected taking part in making the rules, monitoring by the users themselves, graduated sanctions, cheap ways of resolving conflict, and the right to organise recognised by outside authorities. Read beside gainsharing, the list is striking. Clear boundaries is the bonus group drawn around cooperation. Participation in rule-making is the vote that Dong-One Kim found associated with survival.111 Monitoring by the users is the visible measure on the board. The free-rider problem that Eugene Kandel and Edward Lazear described in 1992 is the problem of the commons, and their own proposed answer, peer pressure and mutual monitoring among people who can see each other's work, is close to Ostrom's.112
A firm's improvable processes are a kind of commons. Many people can improve them; none can do it alone; the benefit is shared whether or not the effort is. Gainsharing is one set of rules for governing that commons. Ostrom's work suggests the rules work best when the people who use the resource make them.
Placed beside its neighbours, the limits of gainsharing are easier to state. Each comes with a way through, and each way through leads to one of the companion books.
It cannot change who owns the firm. A plan can be withdrawn by an owner who sells, or by a new management that never believed in it. Santa Anna's book names the fear this creates, writing about workers whose companies are sold:
the fear that the structures you depend on will be captured by interests that don't care about you.113
The way through is to combine the plan with a structure that protects it: an ownership stake, a written charter for the plan that survives a change of control, or the steward-ownership forms the book describes. The companion book on patient capital reads the evidence on how ownership horizons shape what firms keep.
It cannot make an unfair system fair by measuring it. A shared measure in a firm with deep inequities of pay or treatment will make those inequities more visible, not less. Santa Anna is direct about this in his discussion of pay transparency, and the warning carries over intact:
you can't make an unjust system transparent and expect positive results.114
The way through is to fix the obvious inequities first, or at least alongside. A plan that opens the books on an unfair base will be read, correctly, as an invitation to work harder inside the unfairness.
It cannot replace the slow work of changing how people relate. Chapter 3 quoted the book's sentence that authority cannot be distributed to nervous systems still organised around fear. The same passage describes the change involved:
The transition from traditional hierarchy to distributed authority is a developmental process, not a structural flip115
Gainsharing installed as a structural flip, a formula announced on a Monday, meets that limit at once. The way through is the sequence Chapters 5 and 8 laid out: the answer first, then the measure, then the committees, then the design and the vote. Each step gives people time to discover that the new structure is safe before the next one asks more of them.
The root, named once
Haute Lumière · The Gain Was Always Shared · Chapter Twelve
There is a reason the same limits recur across all these books. Santa Anna's book locates it below any particular scheme, in the assumptions a firm makes about where its value comes from. Describing how ordinary corporations come to do harm without anyone intending it, it says:
Not because the people running the corporation are evil, but because the metaphysical assumptions embedded in the system make destruction rational.116
Shrink that to the scale of one plant. If a firm assumes that value comes from capital and management, and that labour is a cost to be minimised, then ignoring the night-shift operator's knowledge is rational. Nobody needs to be indifferent to it. The firm's accounts simply have no line for it. Gainsharing, in its original form, was a correction to that assumption written into the accounts: a line that said the people doing the work produce some of the gain and will share in it. The formula is the visible part. The correction is the substance.
The book puts the same thought more briefly, in its diagnosis of what goes wrong in economic life:
The wound isn't greed. Greed is a symptom.117
A leader who reads gainsharing as a way to buy effort is treating a symptom. A leader who reads it as a correction to the firm's account of where its value comes from is working at the root.
None of what follows requires a change in anyone's pay. It is the first half of Scanlon's idea, which Chapter 2 argued should always come first. The field manual in Chapter 8 is the long form; this is the short form, for a leader who wants to begin before finishing the book. It can begin on Monday and be judged in ninety days.
At the start of Monday's team meeting, speak first, honestly. Santa Anna's book describes a protocol in which the leader opens a weekly meeting not with a status update but with an account of what they are struggling with and what they do not know. It estimates the cost precisely:
This takes roughly ninety seconds and changes the entire relational field of the meeting.118
The book's claim about the effect is the author's, not a measured result. The cost is real and small. A leader who wants ideas to move upward has to show that saying something uncertain is safe.
Announce one rule. Every idea about cost, quality, safety or waste receives a written answer within ten working days, with a reason if it is declined. Say who will answer, and say where the answers will be posted. This is the one commitment in the list that costs a leader something every week. It is also the one the whole book has pointed to as the plainest test of whether a firm means it.
Choose one number and put it where people can see it. A number the team can move by changing how it works. Defects, downtime, rework, waiting time, errors returned by the next team. Post it weekly. When an idea is implemented, write the name of the person who had it beside the line it changed.
Start counting. Ideas received, from how many people, median days to answer, share implemented. These are the vital signs of the circulation. The First-Year Ledger in Chapter 5 is one way to keep them, and Chapter 9 explains how to read them.
Write down what you expect. Before anything is shared, write what you think will change, by how much and by when. In ninety days, compare. That page is your firm's own study, and it will tell you more about your people than any figure in this book.
Talk to the supervisors before you talk to anyone else. Chapter 3 named the supervisor as the person a participation system most often threatens. Before the rule is announced, sit with each supervisor and say what their part will be: the person who brings the team's ideas forward and is named when they are adopted, not the person whose omissions the cards expose. Ask what they would need for that to be true. Measure them, for the first ninety days, on the share of their team's ideas answered on time and nothing else.
Decide what happens to a job an idea removes. Sooner or later someone will suggest a change that makes a task unnecessary. Decide now, and say aloud, what happens to the person whose task it was. The plant in Michael Schuster's study of up to nine years kept employment and turnover stable while productivity rose.119 A firm that cannot make some version of that promise will find that the ideas it receives are always about someone else's work.
Leave the formula until last. If, after six months to a year, the ideas are coming, the answers are going back, and the number is moving, then design a shared reward with the people it would cover, and let them vote on it. If they are not, no formula will fix it.
The living-systems reading of the whole
Haute Lumière · The Gain Was Always Shared · Chapter Twelve
Step back from the plant, the hospital and the household, and the whole of this book can be read as one image. Ecstatic Economics gives it, in a passage about what regenerative enterprise might learn from the natural world. A forest, it observes, does not merely sustain itself; it builds soil, holds water and increases the life it carries over time. Then it says what that means for a living system:
A healthy ecosystem is not in equilibrium.120
The passage continues by naming the condition it calls generative surplus, a system
continuously producing more life than it consumes120
and it lists the design principles a human structure might learn from that, ending with
the distribution of intelligence to every node in the network120
That last phrase is gainsharing in ten words. A firm in which intelligence sits only at the top is a forest in which only the canopy can feel the weather. Every root that meets a stone, every leaf that meets a shadow, holds information the whole tree needs, and a tree that cannot move that information dies of what its own leaves knew. The suggestion card, the committee, the answer within ten days, the number on the board and the share returned when it moves are the vascular system of an organisation that has decided to let its intelligence travel.
Read this way, the gain is not something the firm creates and then decides whether to share. It is the surplus a living system produces when its circulation runs: the fewer defects, the less downtime, the better handover, the report nobody needed that stopped being written. It was produced by many people. Sharing it is not generosity; it is accurate bookkeeping. And the plan's survival depends, as the forest's does, less on any single season's yield than on whether the channels stay open when a season is poor.
The book's closing claim is therefore modest and, the evidence suggests, durable. An organisation that asks, answers, shows and returns is healthier than one that does not, in ways that can be counted, and the first three of those four cost almost nothing to begin.
These chapters have stayed with the published record and with practice that any leader can begin. The finer work of applying them inside a particular organisation, choosing the measure, drawing the group, sequencing the first year and building the structure that lets a plan outlast a change of management, is the subject of the consulting conversations Ammanuel Santa Anna holds with leaders, which begin at hautelumiere.life.
The lens behind the book is set out in full in Ecstatic Economics, in its colour art edition at Ecstatic Economics, and the companion books in the series are listed at https://hautelumiere.life. Two Alpha Papers sit closest to the Monday list: Alpha OKR for keeping the chosen measure tied to the firm's strategy, and Alpha Stakeholder Map for seeing who gains and loses voice when the rules change.
Go back to the night shift, to the operator and the sealing jaw. On Monday nothing about her pay has changed. What has changed is that she wrote the card, someone answered it in writing within ten days, the jaw was reset on a four-hour schedule, and the defect line for that machine dropped by the end of the month. Her name is beside it on the board in pencil.
She is a composite, from the plant in Chapter 5, and so is the card. But the sequence is the one the record keeps finding. That is all gainsharing is, before the money: a firm that has started to write down where its gains come from. Scanlon began there in a failing mill in the late 1930s. Every plan that has worked since has begun there too, whether or not its designers knew the history. The cheque, when it comes, only confirms what the card on the board already says.
One page, filled in on the first Monday and again at day ninety. It covers the first half of the idea only, the part that needs no change in pay.
Day one
Team or unit: · people: ___ · leader:
Monday meeting: leader spoke first about something uncertain? yes / no · what:
The answer rule, as announced: every idea answered in writing within ___ working days, with a reason if declined. Who answers: · Where answers are posted:
The one number: · today's value: ___ · where it is posted: · how often:
Weekly (one line each week)
| Week | Ideas in | From how many people | Answered on time | Implemented | The number | Name written on the board |
|---|---|---|---|---|---|---|
| 1 | ||||||
| 2 | ||||||
| 3 | ||||||
| … | ||||||
| 13 |
Day ninety
Actual against expected, line by line:
One change for the next ninety days:
Inequities that the visible number has exposed, and what will be done about them:
Is the team ready to help design a shared reward and vote on it? yes / not yet · reason:
The best idea of the ninety days, in the words of the person who had it:
Appendix
The Workbook
Haute Lumière · The Gain Was Always Shared · Appendix
From Chapter 1 · Who Holds What a Firm Learns.
Use this before reading further, alone or with two colleagues. It takes twenty minutes. It is not a survey of your people; it is a reading of your structure. Keep the page; the ninety-day plan in the Appendix returns to it.
1 · What is known and not said
Three things the people doing this work probably know about cost, quality, waste or safety that have not reached anyone who could act on them: · ·
How you know, or why you suspect it:
2 · The routes
Every route by which an idea from the floor can reach a decision today (a person, a meeting, a form, a channel):
For each route, who stands at the gate:
Which route was used most recently, and when:
3 · The last time
The last idea from someone without authority that you know reached a decision:
What happened to the person who offered it:
4 · Why the card stays in the pocket
Tick the reasons that apply here: no route · no answer comes back · speaking has cost people before · the idea would expose a manager · the idea might remove a job · nobody can see whether ideas change anything · other:
5 · The first promise you could keep
One promise about answering ideas that you could keep from next Monday without anyone's permission:
From Chapter 2 · The Mill That Asked First.
Use this before any conversation about shared rewards. It takes one hour, with the leader and two people who do the work. Write the answers down; they become the baseline the later tools in this book return to.
1 · The question we are asking
What do the people doing this work know about cost, waste or quality that management does not? Write three examples, in their words:
When was the last time one of those examples reached a decision-maker? · What happened to it?
2 · The answer we owe
How is an idea submitted today? (none / informal / a form / a committee)
Median days from submission to a written reply, measured from the last ten ideas:
The promise we can keep from Monday: every idea answered in writing within ___ days, with the reason if declined. Owner:
3 · The measure people can see
One number this group could move by changing how it works:
Three things that move it which this group does not control:
Can the people in the group see this number today? yes / no · How often?
What would have to be opened (a report, a ledger, a meeting) for them to check it?
4 · The group
Who cooperates to produce a change in that number? List the roles:
Who is left out of that list but affects the number?
Is the group drawn around cooperation or around the org chart?
5 · Before any money
If fewer than four are ticked, the next step is the unticked item, not a pay scheme. Write it:
Review date, ninety days from today:
From Chapter 3 · Where the Gain Goes Missing.
Run this before any shared-reward plan is announced. Imagine it is two years from now and the plan has been abandoned. For each critique, write the most likely reason it failed, then the step that would prevent it. One hour, with the leader, a supervisor, and two people the plan would cover.
1 · One share among many (free riding)
Where the group cannot see itself, the visibility we will build (a board, a weekly review, a naming habit):
Smallest group that still contains the cooperation that produces the gain:
2 · The ratchet and the base
Proposed base period or standard: · Was it a normal period? yes / no · evidence:
Written rule for when and how the base changes:
Will the people covered vote on the plan and the rule? yes / no · if no, why not:
How capital investment will be treated:
3 · Power and the supervisor
Each supervisor's role in the plan, in one sentence:
How supervisors will be credited when their team's ideas are adopted:
Written-answer rule: every idea answered within ___ days, with a reason if declined · Who checks it is kept:
The first sign that power has not moved, and who will report it:
4 · The number moves for other reasons
Things inside the group's control that move the measure:
Things outside it that move the measure:
How each outside event will be adjusted for, decided now:
When the easy gains are exhausted, what training will widen what people can change?
5 · The evidence sees the survivors
What we expect to change, by how much, by what date (written before the plan starts):
Where we will record what fails, and who keeps it:
The verdict
Critiques with no preventing step written:
If any, the plan waits until each has one. Review date:
From Chapter 4 · A Firm Has a Circulation.
Draw this on one page for any team or unit, with or without a gainsharing plan. Fill it from records where they exist and from honest estimates where they do not, and mark which is which. Repeat it every quarter.
1 · The up-current (knowledge moving towards the people who can change the system)
Routes by which an idea can travel today:
2 · The answer (the first return)
3 · The measure (the second return)
The number this team can see that tells it whether its ideas changed anything:
How often the team sees it: · Who shows it to them:
4 · The share (the third return)
What, if anything, returns to the team when the measure improves (recognition, time, training, money):
Has the target ever been raised because the team beat it? yes / no · when:
5 · Where the current blocks
The one change that would reopen that stage this quarter: · Owner:
The vital sign that will show whether it worked, and its value today:
6 · The soil
The most recent example of someone saying so, and what happened next:
From Chapter 5 · Twenty-One Defects in a Thousand.
Keep this as one page per month for the first twelve months of any participation system. Every entry is a count or a date, not an opinion. It is the record that lets a firm become its own case study.
Header
Unit: · People covered: ___ · Month ___ of 12 · Prepared by:
1 · The up-current
2 · The answer
3 · Implementation
4 · The measure (collected before anything is shared)
Measure chosen: · this month: ___ · same month last year: ___
Outside events this month that moved the measure, and by roughly how much:
5 · Employment
Any role removed because of an implemented idea? yes / no · if yes, where the person went:
6 · Structure
Outside help in use: none / building / operating · planned end date for outside help:
7 · One line
The most useful idea this month, in the words of the person who had it:
From Chapter 6 · Talk Is the First Return.
Use this for the first six months of any participation system in service work, before the shared measure has had time to move. Keep one page per month. Every entry is a count, a date or a quotation, not an opinion.
Header
Team or unit: · People covered: ___ · Month ___ of 6 · Comparison team (if any, and why it is comparable):
1 · Talk about the work
2 · Ideas moving between people
One example, in the words of the person who adopted it:
3 · Noticing and advising
The shared phrase this team uses for an advisory remark:
4 · The rating
Any case this month where someone who made an advisory remark was rated down for it: yes / no · what was done:
5 · The measure (watched, not yet judged)
Outcome measure chosen: · this month: ___ · same month last year: ___
Written expectation for month six, set before month one:
6 · What survives
Parts of the participation system written into how the team works, independent of any bonus: meeting / answer rule / visible measure / naming who found what / other:
If the reward ended tomorrow, which of these would continue?
From Chapter 7 · Promises, Prices and the Operating Room.
Use this before any shared-gain arrangement begins, and monthly once it is running. It has two halves: the promises the arrangement makes, and the sources of whatever gain it records.
Header
Organisation or unit: · People covered: ___ · Month ___ · Prepared by:
1 · The purpose and its guardrail
What this organisation exists to protect, in one sentence:
The guardrail measure(s): · this month: ___ · baseline: ___
Written rule, agreed before the start: if the guardrail falls, the gain measure does not count. Agreed on (date): · by whom:
2 · The promises
| Promise made | To whom | Kept this month? yes / no | If no, the reason given, and by whom |
|---|---|---|---|
| Every idea answered in writing within ___ days | |||
| The measure visible to everyone covered | |||
| The rules for changing the base unchanged without notice and a vote | |||
| No role removed by an implemented idea without (stated alternative) | |||
| Other: ______ |
Was the arrangement offered in exchange for a concession? yes / no · If yes, how the participation was made real before the first payment:
3 · The sources of the gain
Total movement in the gain measure this period:
Of which, from changed practice (what people did differently):
Of which, from changed prices (what the organisation paid):
Of which, from events outside the group's control:
Shown to the people covered? yes / no · how:
4 · Power
Who decided the most consequential change this month, and were the people covered part of the decision?
5 · Advice
Legal or regulatory review completed (where the field requires it): yes / no / not applicable · date:
6 · The record
The most important promise kept this month:
Any promise broken, and what was done about it:
From Chapter 8 · The Field Manual.
One sheet, kept by the unit leader and the secretary together, reviewed at the end of each phase. Every line is a date, a name or a count.
Roles (named in week one)
Sponsor: · committed until (date, at least two years):
Unit leader:
Supervisors:
Union officers, if any:
Secretary (elected week nine): · protected hours per week: ___
Finance person:
Outside help, if any: · leaving date written in their terms:
Phase one · Readiness (weeks 1–4)
Origin Reading done (date): · with whom:
Operational data available, measure by measure, and months of history:
Pre-mortem done (date): · critiques without a preventing step:
Sponsor's four answers: two-year commitment yes / no · promise on removed jobs (in words): · financially sound yes / no · known inequities and the date they will be fixed:
Decision to begin: yes / not yet · if not yet, the reason and the work for the next ninety days:
Phase two · Supervisors first (weeks 5–8)
Union officers met (date): · what they asked for:
Phase three · The answer rule (weeks 9–22)
Announcement date: · the five points said by shift: yes / no
Answer window: ___ working days · where answers are posted:
Phase four · The measure (weeks 23–40)
Measure chosen: · who counts it: · books opened to all (date):
Outside events that move it, and how each will be noted:
Sealed expectation filed (date): · where:
Phase five · Committees (weeks 41–52)
Elections held (date): · committees formed: ___
Committee budget limit per idea, set by sponsor:
Outside help left on (date): · as planned yes / no
First annual review (date): · expectation opened and compared yes / no · posted yes / no
Phase six · Design and vote (months 13–15)
Design group members:
Sponsor's statement on affordability, before ballot (date):
Ballot date: · threshold set in advance: · result:
Phase seven · Review and plateau (months 16–18)
First review (date): · rules kept yes / no · employment held yes / no · roles removed by ideas and where each person went:
Boundary training started (date): · committee rotation yes / no
Change of control
Charter adopted (date): · held jointly by:
Record stored at:
Ownership options considered (date, and by whom):
From Chapter 9 · How to Know It Worked.
A quarterly review, two hours, conducted by the screening committee with the sponsor. Fill every line from the secretary's ledger and the posted measure. Where a figure is an estimate, mark it as one.
1 · The before
Was the measure already improving before anything began? yes / no / unclear · evidence:
Sealed expectation opened (date filed):
2 · The vital signs this quarter
3 · The work
Did the line step at any phase? which: · by more than its ordinary month-to-month variation? yes / no
Did the trend change after any phase?
4 · The people
Roles removed by implemented ideas ___ · where each person went:
5 · The comparison
Comparison unit, if any: · same measures collected from the same date? yes / no
Change in the comparison unit over the same period:
6 · The survivors
Leavers asked about the system: ___ of ___ · said it played a part: ___ · what they said:
Interruptions this quarter (committee missed, measure not posted, answers late), with dates:
7 · Expected against actual
| Measure | Expected (sealed) | Actual | Gap | Why |
|---|---|---|---|---|
| Ideas per month | ||||
| Median days to answer | ||||
| Share implemented | ||||
| The chosen measure | ||||
| Voluntary turnover |
8 · Where the circulation blocks
One change for next quarter: · owner:
9 · What this reading cannot show
In one paragraph, the alternative explanations this quarter's evidence cannot rule out:
From Chapter 10 · The Hard Conversations.
Use one sheet for each hard conversation, filled before the meeting and completed after it. Keep the sheets together; they are part of the record Chapter 9 asks a firm to keep.
Before
Who: · role: board / owner / union / supervisor / finance / partner / family / sceptic
What this person is most likely to fear, in their terms, not ours:
What the evidence can honestly say to that fear, and its limit:
What we are asking them to decide, in one sentence:
What we are not asking for (write it, so it can be said aloud):
The date by which the decision would be reviewed:
Any figure we plan to mention, and where it was measured: (if none was measured, mention none)
The script, in our own words (five to eight sentences)
What we want:
What we do not know:
What we will show them, and how often:
What happens to a job an idea removes (if relevant):
The invitation into the work:
During
Their first concern, in their words:
What they would need for a yes:
Anything they told us that changes the plan:
After
Decision: yes / not yet / no · conditions:
What we committed to, and by when: · owner:
Next conversation with this person:
If not yet or no: the part of the plan we can still do without their approval:
From Chapter 11 · One Person's Gain.
One page, for a household or a person working alone. Fill it on the first Sunday of a month and again at the end of the third month. It involves no financial products and no change to anyone's income.
1 · The people
Who shares this budget or this work:
For each person, one sentence on what money felt like in the house they grew up in (optional, said aloud rather than written if preferred):
2 · Who knows what
Three things someone here notices about where money or time goes that the person deciding may not:
When each was last mentioned, and what happened:
3 · The answer
Our rule: every suggestion gets a reply within ___ days, even if it is no. Who replies:
4 · The one number
The number: · today: ___ · where it will be posted:
Things we control that move it: · things we do not:
5 · What comes back
If the number improves, what returns, and to whom:
Has the household or the person ever responded to a saving by expecting more? yes / no · when:
6 · The weekly inventory (one line a week)
| Week | Event | State during it (settled / anxious / shut down) | More or less alive afterwards |
|---|---|---|---|
| 1 | |||
| 2 | |||
| … | |||
| 13 |
7 · At the end of three months
The number, then and now:
What the inventory showed that the number did not:
One change for the next three months:
From Chapter 12 · The Gain Was Always Shared.
One page, filled in on the first Monday and again at day ninety. It covers the first half of the idea only, the part that needs no change in pay.
Day one
Team or unit: · people: ___ · leader:
Monday meeting: leader spoke first about something uncertain? yes / no · what:
The answer rule, as announced: every idea answered in writing within ___ working days, with a reason if declined. Who answers: · Where answers are posted:
The one number: · today's value: ___ · where it is posted: · how often:
Weekly (one line each week)
| Week | Ideas in | From how many people | Answered on time | Implemented | The number | Name written on the board |
|---|---|---|---|---|---|---|
| 1 | ||||||
| 2 | ||||||
| 3 | ||||||
| … | ||||||
| 13 |
Day ninety
Actual against expected, line by line:
One change for the next ninety days:
Inequities that the visible number has exposed, and what will be done about them:
Is the team ready to help design a shared reward and vote on it? yes / not yet · reason:
The best idea of the ninety days, in the words of the person who had it:
The Tools above are instruments; this is the order in which to pick them up. It covers the first half of the idea only, the part that needs no change in anyone's pay, and it is written for a unit of between ten and two hundred people. A household can follow the same weeks at its own scale with the Kitchen-Table Circulation in place of the ledgers.
Before day one (one week)
Weeks 1 to 2: the answer
Weeks 3 to 6: the measure
Weeks 7 to 10: the circulation
Weeks 11 to 13: the reading
What not to do in the first ninety days
At day ninety the page will show one of two things. Either ideas are arriving, answers are going back and the number is moving, in which case the organisation has the circulation that every working plan in this book began with. Or one of those three is not happening, in which case the page shows exactly where, and the next ninety days have their work.
Somewhere tonight a person who does the work knows how to make it better and is not going to say it. Seventy years of evidence describe what happens when somebody asks first.
Luminous Prosperity Inc. takes this work into organisations as a gainshare. The flow is opened from the inside, with the people who already hold what the firm knows. The gain is counted from records the firm already keeps, against a measure both sides sign before anyone knows the answer, and the house is paid as a share of what is found: between thirty and ten per cent of it.
The workbook carries a team to its first plan. A whole firm, with a measure both sides will sign and a share that returns, is worked in the building. That begins with a conversation.