Haute Lumière

Commerce · VI.02 · MMXXVI · daylight

La Bourse  /  Volume VI  /  Nº VI.02  /  Workbook — the Gainshare employee

WORKBOOK — THE LUMINOUS GAINSHARE EMPLOYEE

Chapter VI.02 · Ostrom, Fully Applied

For the person working inside a gainshare arrangement — where a defined share of verified improvement returns to the people who created it. A gainshare is a commons: the pool of measured improvement is the resource, the scheme document is the rulebook, and every one of Ostrom's twelve conditions has a line in it. This workbook reads your scheme as a commons, scores it, and tells you what to ask for.

---

YOUR SCHEME IS A COMMONS, AND THAT IS NOT A METAPHOR

A gainshare pool has every property of a common-pool resource in the technical sense.

• It is subtractable at the margin. A share paid to one claim is not available to another, unless the formula is genuinely uncapped — and almost none are.
• Provision is under-supplied relative to appropriation. Improving the measured number is costly to the individual and benefits everyone in the pool. That is the free-rider structure exactly.
• Boundaries decide everything. Who is in the pool, what counts as improvement, and over what period.
• It needs monitors, sanctions and a conflict route, or it degrades into a discretionary bonus wearing the word gainshare.

So the twelve principles are not an analogy you are being offered. They are a line-by-line audit of your scheme document, and the empirical record behind them — 91 studies, 77 cases, and a configurational re-analysis of 69 — tells you which lines matter most.

Two facts to carry from that record before you open the document.

No single condition is sufficient. Baggio and colleagues found no design principle both necessary and sufficient. A scheme with an excellent formula and no conflict route is not four-fifths of a good scheme.

Four conditions carry most of the load. Their absence — 2A rules fitted to conditions, 2B appropriation matched to provision, 4B monitors answerable to the monitored, and 5 graduated sanctions — was most strongly associated with failure. Those are the four to read first.

---

PART ONE — DISCOVERY

Read your scheme as a rulebook

Exercise 1.1 — The twelve-line audit (2 hours)

Take your scheme document and a highlighter. For each of the twelve, find the clause that satisfies it and write the clause number beside it. Where there is no clause, write none.

The count of "none" is your first finding, and it is worth more than any individual answer.

Exercise 1.2 — The baseline question (30 minutes)

Before anything else, answer the question that decides whether a gainshare is worth being in at all:

What happens to the baseline when the gain is realised?

If the baseline resets to the improved level each period, you are on a treadmill. Every gain raises the bar you are measured against, the same effort yields less each cycle, and eventually nothing. This is baseline ratcheting and it is the most common way gainshare schemes quietly die.

In the language of this chapter, ratcheting is a failure of principle 2A — the rules are not congruent with the conditions, because the conditions now include the improvement you made. A well-designed scheme either holds the baseline for a stated term, or ratchets on a published, gradual schedule everyone can see coming.

Find out which yours does. If nobody can tell you, that is the finding, and it is the most valuable thing you will produce this month.

Exercise 1.3 — Find the hay rule (45 minutes)

Every durable commons has a rule that runs on something already visible, so that monitoring costs nothing. Törbel pegged summer cows to winter hay: nobody audits the pasture, everybody can see the barn.

What is already visible in your work that tracks contribution? Shift coverage. Tickets closed and reopened. Rework rate. Handover quality. Whatever it is, it is the candidate for principle 2B, and a formula built on something already visible is a formula nobody has to police.

---

PART TWO — THE ARITHMETIC

Score it, and know what the score is worth

Exercise 2.1 — Score your scheme (90 minutes, ideally with one colleague)

  0  absent
  1  present in form only
  2  present and operating
  3  present, operating, and measurable from outside without permission

That last band is strict and deliberately so. A clause you cannot verify without asking permission scores 2, not 3. If you cannot see the verification working, you are trusting rather than knowing, and the whole point of a written scheme is to replace trusting with knowing.

Report three numbers, never one:

  full score              twelve principles       out of 36
  core score              1B 2A 2B 4B 5 6         out of 18
  non-core score          the other six           out of 18

The worked case in the chapter — English Wikipedia — scored 29 of 36 (80.6 %) overall, with a core of 16 of 18 (88.9 %) and a non-core of 13 of 18 (72.2 %). Its two weak scores both sat outside the load-bearing set, which is why the core read higher than the whole.

Run the same comparison on your scheme. If your core is lower than your whole, you have the opposite and more dangerous shape: a document that looks complete and is weak exactly where the evidence says it matters.

Exercise 2.2 — What the score is worth (45 minutes)

You will be tempted to read a probability off your total. Do not, and here is the arithmetic for why.

The evidential weight of a condition is a likelihood ratio — how much more often it is present in schemes that last than in schemes that die — expressed in bits:

     evidence  =  sum over i of  (score_i / 3) x log2(LR_i)

For a score of 29, the sum of score/3 is 9.667. Nobody has measured the likelihood ratios. Sweep them, from prior odds of 0.4658:

  LR = 1.2        2.54 bits       posterior  73.08 %
  LR = 2.0        9.67 bits       posterior  99.74 %
  LR = 4.0       19.33 bits       posterior  99.9997 %

A band 16.79 bits wide. The score is a diagnosis — it tells you exactly which conditions are unmet — and it is not a forecast.

This matters to you practically. When somebody presents the scheme's design as evidence that it will last, the honest answer is that the design tells you where it is weak, and the thing that tells you whether it lasts is whether anybody has ever watched a scheme like it die. Ask whether they have.

Exercise 2.3 — Count the dead schemes (one week)

This is the exercise that will change how you negotiate.

Find out how many gainshare or profit-share schemes this organisation has run before, when each started, and when each ended. Ask HR, ask long-serving colleagues, look in old handbooks.

You are building a base rate, and the reason is that almost nobody has one. The only comparable population census in the digital commons literature — Schweik and English on the whole of SourceForge — found:

  projects, 2009 census                    174,333
  success in growth                         24,899      14.28 %
  abandoned in growth                       53,450      30.66 %
  determinate growth outcomes               78,349
  success share                                         31.78 %

Roughly two in three died. And the wiki literature studied 740 wikis — openly, the top one per cent — implying about 74,000 in the population and 73,260 nobody ever looked at.

Your organisation's schedule of past schemes will be short, and it will be the only base rate anybody in the building has. That makes you the person with the denominator, which is a considerably stronger position in a negotiation than being the person with the opinion.

---

PART THREE — WHAT TO ASK FOR

Five asks, in order of how much they cost the company and how much they protect you. Every one of them is a principle, and naming the principle is what turns a complaint into a proposal.

1. A baseline term — principle 2A. Costs nothing.

Ask for the baseline to be held for a stated term, or ratcheted on a published schedule. This costs the company no money in the current period and removes the single most common cause of scheme death. It is the cheapest ask on the list and the one with the highest return.

2. Visible contribution — principle 4A. Costs a report.

Ask to see the components of the pool, not just the total. If the data exists for the calculation, it exists for the report. This is the cheapest gap to close in almost any scheme, and the reason is structural: in a well-instrumented workplace, monitoring is a byproduct of working. The scale of that advantage is worth knowing — Wikipedia holds its entire enforcement record, every edit ever made with the diff showing what changed, for:

  edits, lifetime                    1,370,633,170
  at 500 bytes per revision delta            685.3   GB
  at $0.023 per GB per month                $189.15   / yr

Your scheme's transparency costs the same order of nothing. Ask for it.

3. An accountable verifier — principle 4B. Costs a line in the deed.

Ask who verifies and to whom they answer. If the verifier is the same function that pays the pool, the principle fails. The record is unusually clear here: accountable monitoring was the most strongly supported finding in the whole Cox meta-analysis. The ask is a rotating employee representative in the verification, or an external practitioner with a published scope.

4. A conflict route with a clock — principle 6. Costs a service level.

Ask where a disputed calculation goes, how long it takes, and what it costs you to raise. Then ask for the median resolution time to be published. The Valencian water tribunal has run for centuries on one property above all: the cost of bringing a complaint is low enough that complaints get brought early, while they are still small. A dispute route with an unpublished median is one nobody trusts until it is too late.

5. Notice on unilateral change — principle 7. Costs a clause.

Ask what happens if the company changes the scheme. You want enumerated grounds, a stated notice period, a written reason, and a route to the conflict panel. This is the ask people skip because it feels like asking for distrust. It is not. It is the clause that makes the other four durable, and on 10 June 2019 the largest digital commons in the world discovered exactly what it cost not to have it — administrators, bureaucrats and years of goodwill, over a power everybody had forgotten was held.

---

PART FOUR — HOW THE LEDGER WORKS, READ FROM INSIDE

Four things to be able to state about your own pool without looking them up.

The formula, as an equation. Not "a share of improvement" — the actual expression, with the baseline term, the period, the cap if there is one, and the treatment of costs. If you cannot write it down, you cannot check it, and a share you cannot check is a promise.

The verification date and the payment date. The gap between them is where schemes lose trust. Know both, and know who decides if verification slips.

Your own contribution, in the units of the formula. This is the discipline that turns you from a beneficiary into a participant. Track it weekly. It takes minutes and it is the single thing that makes a conversation about your share concrete rather than plaintive.

What the pool did last period, and the period before. Two numbers. If they are not published, that is a principle 4A gap and Ask 2 covers it.

---

THE STANDING PRACTICE

Once a quarter, one hour:

1. Re-score the twelve. Note every change and what caused it.
2. Check the baseline — has it moved, and was the movement on the published schedule?
3. Add any scheme that ended this quarter, anywhere in the organisation, to your base-rate schedule.
4. Take one ask from the list of five and move it forward by one step.

Four hours a year. At the end of three years you will have the scored history of your own scheme, a base rate nobody else in the building has, and a record of which asks landed — which is, in the language of this chapter, an inception cohort of one, coded blind, followed to outcome.

That is a smaller version of the study the whole field is missing. It is also the best possible preparation for the conversation in which somebody hands you a scorecard and reads a probability off it.

Ask them the question:

What is the likelihood ratio, and who measured it?

You will know whether they have one. Almost nobody does, and the score is still worth having — it is worth far more when nobody is pretending it is something else.

And hold on to the reason this is worth four hours a year rather than an afternoon's irritation. A gainshare that works is one of the few arrangements in commercial life where the interests of the people doing the work and the interests of the organisation are pointed in the same direction by a written formula rather than by an appeal to goodwill. That is rare and it is worth defending carefully. Every one of the five asks above makes the arrangement more durable for the company as well as for you: a scheme that nobody trusts is a cost with no behavioural return, and a scheme that ratchets its own baseline into irrelevance was an expense the company paid for twice.

So none of these conversations is adversarial, and none of them needs to be framed that way. You are asking for the conditions under which the thing you are both already trying to do keeps working — and you are asking with the empirical record of thirty-five years of commons research behind you, which is a better position than most people bring to a conversation about pay.