First Principles (10)

Inherited structures are rebuilt from what they are for, not preserved because they were inherited.

Trust Architecture · Pillar III · Reward

The definition

The condition. Reward structures are periodically tested against the behaviour they exist to produce, and retained, redesigned, or retired on that basis rather than preserved because they were inherited.

What it means. The organization periodically returns to the purpose of each incentive, metric, promotion path, and status marker, and tests whether it still produces the behaviour it was built to produce. Where it does, it stays. Where it no longer does, it is redesigned or retired, rather than preserved merely because it already exists.

The failure it prevents. A reward system that once tracked real contribution decays back into rewarding ritual and incumbency. Without periodic purpose-testing, incentives drift, metrics harden into targets gamed for their own sake, and yesterday’s evidence of merit becomes today’s empty performance, while the structures persist because changing them is costly and keeping them is free.

How it holds truth against power. Reward structures can become shelters for old power, because their existence arrives pre-justified. An incentive or status marker may persist through simple inertia, but where it keeps distributing status, authority, or advantage, those who benefit can defend that distribution without defending it directly. “This is how we have always measured it” turns a contestable choice into an apparently neutral fact. First Principles reverses the move by requiring the reward structure to justify itself through its present purpose and effects.

Connections. The Reward pillar’s safeguard over time: it keeps Substance Over Signal (8) and Costly Action (9) from ossifying, so the system keeps rewarding real contribution and real commitment rather than the rituals that once signalled them. The pillar-scale instance of the stance that Structural Reform (12) applies to the whole architecture. Re-tests, within the reward system, whether word and deed still align (the Foundation, 0).

The test. Which inherited reward structure here is now preserving status, ritual, or incumbency rather than rewarding the conduct it was created to produce?

In depth

Somewhere in most organizations is a reward structure that no longer does what it was built to do. A metric everyone has learned to hit without producing the thing the metric was a proxy for. A promotion path that once tracked contribution and now tracks tenure, or visibility, or simply time served. A bonus formula, an award, a ranking, a title, designed years ago to reward one behaviour and now quietly rewarding another, or nothing at all. Ask why it works this way and the answer, eventually, is a version of: this is how we do it here.

That answer is where this principle begins. A reward structure that has drifted from its purpose has not become harmless. It keeps paying out, and what it pays for is now something other than what anyone intended.

What it is, and what it isn’t

First Principles, in the Reward pillar, is the principle that reward structures are periodically tested against the behaviour they exist to produce, and retained, redesigned, or retired on that basis, rather than preserved because they were inherited. The organization returns, on purpose and on a schedule, to a plain question about each structure that shapes who gets rewarded: what is this for, and is it still producing that?

Notice first what it is not. It is not change for its own sake, and it is emphatically not a presumption that inherited structures should be torn down. A great many reward structures still do exactly what they were built to do, and purpose-testing will examine them, confirm it, and keep them unchanged. Testing against purpose is not the same as replacing; a structure examined and found sound has passed the test, not failed it. The target is never age. It is the reward structure whose only remaining justification is that it already exists.

Nor is it a single act of redesign. It is a habit, a recurring return to purpose, because the failure it addresses is slow. Even a reward system built with great care drifts, as the world around it changes and the behaviours it once tracked stop being the behaviours it now rewards. The line the principle draws is this: when you ask a reward structure what it is for, does it have a living answer that its present effects actually confirm, or only the answer that this is how things are done? First Principles is the willingness to keep asking, and to redesign or retire when the honest answer has died.

It is worth being clear about one boundary. This is the Reward-pillar member of a family. The same purpose-testing stance, applied to the whole architecture rather than to reward structures specifically, is Structural Reform, one of the Bindings. Here the scope is deliberately narrow: the incentives, metrics, ladders, and honours through which an organization decides whom to reward.

The mechanism: why it matters

The engine of the failure is a plain asymmetry between preserving a reward structure and justifying it.

Every reward structure exists to encourage, recognize, or select for some behaviour: to reward contribution, surface strong performers, or make valued conduct pay. Over time the world moves, the behaviour that matters shifts, and people learn the structure well enough to satisfy it without producing the thing it was a proxy for. The structure, meanwhile, stays exactly where it is, because dismantling or redesigning it takes evidence, authority, and effort, while leaving it alone takes nothing at all. Continuity gets a standing discount. An existing arrangement persists by default; a change has to be argued for.

That asymmetry is enough, on its own, to let obsolete reward structures accumulate. Sometimes no one in particular benefits; the structure simply survives through inertia, switching costs, and the fact that questioning it is nobody’s job. But a second force often runs alongside the first. Where a reward structure continues to distribute status, authority, or advantage, the people it advantages have every reason to keep it, and they can defend it without ever naming the advantage, because the structure comes pre-justified by its own existence. “This is how we have always measured it” converts a contestable choice into an apparently neutral fact and closes the question before it opens. Inertia keeps the structure alive; power, where it is present, makes sure the structure is never asked to justify itself. The two reinforce each other.

Stated as failure economics: when a system stops testing its reward structures against purpose, continuity acquires an advantage over justification. Existing arrangements persist because changing them requires evidence, authority, and effort while preserving them requires almost nothing, and where those arrangements also protect incumbent advantage, inertia and power reinforce one another. First Principles is the deliberate correction of that tilt. It puts the burden back on the structure to show, from its present effects, that it still does what it was built to do.

How it breaks

First Principles fails quietly, through the absence of a habit rather than the presence of a villain:

“This is how we measure it.” The reward metric or ranking is justified entirely by its own existence, its original purpose long unexamined and its present effects never measured against that purpose.
The gamed proxy. A metric that was once a good stand-in for a valued behaviour becomes a target pursued for itself, hit reliably while the underlying behaviour it was meant to encourage quietly disappears.
The ossified ladder. A promotion or status structure that once tracked contribution now tracks tenure, proximity, or fluency, and keeps conferring standing on the basis of what used to be evidence.
Incumbent-protected structure. A reward arrangement defended most firmly by those it advantages, its language of purpose serving as cover for its function of preserving their position.

The long view

For decades, one of the most admired reward structures in management was forced ranking. Popularized at General Electric under Jack Welch in the 1980s, the model sorted employees into fixed tiers, a top band to be rewarded generously, a large middle, and a bottom band to be managed out, on the theory that ranking people against one another would identify and reward the strongest performers and steadily lift the quality of the workforce. It had a clear purpose, and for a time it was widely copied, including at Microsoft, which ran a version of it for years.

Then the structure’s actual effects came into view, and they were not the ones it was built to produce. Because a fixed share had to be rated at the bottom regardless of real performance, managers were forced to mark genuinely good people as underperformers to satisfy the distribution. Worse, employees learned what the structure truly rewarded. Former Microsoft executives and employees described engineers avoiding teams that contained the strongest colleagues, because being ranked beside them could damage their own standing, and a system meant to reward performance began quietly rewarding self-preservation and internal politics instead. The reward structure had drifted a long way from its purpose. It was still faithfully producing rankings; it was no longer producing the behaviour the rankings were for.

In 2013 Microsoft retired stack ranking. Its new system removed both the predetermined distribution and employee ratings, placing greater emphasis on teamwork, contribution to others, and business impact. The structure was no longer defended by what it had once promised; it was judged by whether its present effects supported the behaviour Microsoft now needed. Notice what did and did not happen. This was not change for novelty, and it was not a verdict that all ranking is worthless. It was a reward structure tested against its purpose and found to be producing close to the opposite of it. The lesson is the principle exactly: a reward structure preserved because it is familiar, and never re-tested against the behaviour it exists to produce, will keep paying out long after it has begun paying for the wrong thing. What breaks that spell is not a better slogan but a return to purpose, someone with the standing to ask what the structure is now producing, and the willingness to redesign or retire it when the answer comes back wrong.

Where it sits in the architecture

First Principles closes the Reward pillar, and it is the pillar’s safeguard across time.

It keeps Substance Over Signal (8) and Costly Action (9) from ossifying. Those principles tune the reward system to real contribution and to costly, credible commitment. But neither stays tuned on its own. What counts as “substance” and what registers as “cost” gradually harden into fixed proxies, and a proxy, left unexamined, becomes a ritual that can be performed without the thing it once measured. First Principles is what periodically re-tests those proxies against the behaviour they were meant to capture, so that yesterday’s honest evidence of merit does not become tomorrow’s theatre.

It is the pillar-scale instance of Structural Reform (12). The same purpose-testing stance, applied to the whole architecture rather than to reward structures specifically, is one of the Bindings that keeps the entire system alive over time. First Principles is that stance operating within a single pillar; Structural Reform is it operating over everything. They are one discipline at two scales.

And it re-tests the Foundation (0) within the reward system. By asking whether a reward structure still produces the behaviour it was built for, it re-checks whether word and deed still align there, whether the thing the organization claims to reward is the thing it actually rewards. It closes the loop back to the base condition, inside the domain of who gets paid, promoted, and honoured.

Together with its neighbours it completes Reward, the pillar that asks whether the system, over time, keeps the people who tell the truth. It can only keep doing so if the structures that reward them are themselves periodically re-tested against purpose, rather than left to drift back toward rewarding signal, ritual, and incumbency.

How to build it, and test it

To find the reward structures running on inertia, take the incentives, metrics, ladders, and honours that actually shape behaviour, and apply the test to each:

Which inherited reward structure here is now preserving status, ritual, or incumbency rather than rewarding the conduct it was created to produce?

Wherever the honest defence is “this is how we measure it,” and the structure’s present effects are not measured against its purpose, you have found a reward structure running on momentum, and possibly an incumbent who is comfortable with it.

The fixes are structural:

Re-test reward structures against purpose, on a schedule. For the incentives and metrics that matter, require a living answer to what behaviour they exist to produce, and check the structure’s actual effects against that answer, not against its reputation.
Watch for the gamed proxy. Where a metric is hit reliably while the behaviour it was meant to encourage is fading, the proxy has become the target, and it is time to redesign or retire it.
Confront the incumbent case honestly. Where a reward structure is defended hardest by those it advantages, suspect that its language of purpose is covering its function of preserving their position, and re-open the question anyway.
Institutionalize the habit. Make “what is this reward structure producing, and is that what it is for?” a scheduled, owned act, so that renewal does not depend on the arrival of a rare reformer but is built into how the reward system is governed.

One honest caveat: purpose-testing is not a licence for constant churn, and it does not discard accumulated wisdom. Many reward structures encode hard-won lessons and will pass the test easily, and should be left alone when they do. The discipline is to ask, not to assume the answer is no. And when a structure is retired, the people who prospered under it, and the real interests bound up in it, deserve to be dealt with honestly rather than waved away as mere incumbency. What the principle forbids is not old reward structures, but unexamined ones: arrangements that have quietly stopped being asked what behaviour they now produce.

The point

First Principles is the habit that keeps the reward system honest over time, because it re-opens the one question an inherited structure exists to close: what is this for, and is it still producing that? Ask it regularly and the reward structures that still work survive easily, while the ones that survive only on inertia, and the advantage hiding inside them, are finally exposed. Stop asking it, and the standing discount that continuity enjoys will let even a well-built reward system drift, quietly, into paying for ritual, position, and the performance of merit, until the thing you reward and the thing you meant to reward have almost nothing left in common.