Why it matters

Never ask why someone is behaving the way they are until you know what they’re rewarded for. Most of what looks like a people problem is an incentive problem wearing a disguise.

For example: a call center pays its agents by calls handled per hour, and quality collapses — agents rush people off the line, leave problems half-solved, and customers call back three times for one issue. It’s tempting to conclude the agents are careless or poorly trained. They’re not. They’re doing exactly what they’re paid to do. Change the pay to reward first-call resolution and satisfaction, and the behavior flips within a week — no speeches about service quality required. The agents never changed. The reward did.

  • What it reveals. The reward structure actually driving behavior — financial, but also status, approval, autonomy, fear of sanction — as opposed to the character or intentions of the people in the system.
  • How it changes the read. You stop asking “why are these people acting this way?” and start asking “what is this system actually paying them to do?” — and the puzzling behavior turns out to be perfectly rational.
  • When to foreground it. Whenever a system produces outcomes its designers didn’t intend and can’t explain — people gaming a metric, optimizing for an obvious proxy, visibly ignoring stated goals.
  • What you’d miss without it. That exhortation, training, and culture work can’t override a reward structure that points the other way — and that the fix is almost always to change the structure, not the people.
  • Where it misleads. Map only financial incentives and you’ll miss the operative one; and bolt on a new metric without modeling how people will game it, and the second-order response can swamp the first-order gain.

Realtime examples

See real, dated analyses where this discipline shaped the read on the news → Incentives on Main Street Independent

How to invoke it in Ora

You’re trying to understand a situation with several parties pulling in different directions, and you want to see what each one is actually playing for — not the role on their business card, but the reward they’re really chasing.

Describe the situation and the parties, and ask:

“Map the stakeholders in this rollout — who’s involved, what each one actually wants and could lose, and what the system is really rewarding each of them to do.”

Incentives is one of the always-loaded reasoning tools in the Stakeholder Mapping analysis. As Ora inventories the parties and their stakes, this lens re-anchors each party’s behavior onto the reward structure it actually faces — surfacing where what a party is paid (in money, status, or safety) diverges from what everyone says the goal is.

One thing to know: the words stakeholder map, stakeholder analysis, who needs to be at the table, who has standing, or naming the Bryson power-interest grid / Mitchell-Agle-Wood salience are what route you here. The analysis is the map; the incentive lens is what keeps the stakes honest.

Name what each party is rewarded for, not just their title. “The regulator” is a role; “the regulator, whose staff are judged on visible enforcement actions, not on quiet compliance” is an incentive — and only the second predicts behavior.

One thing Ora won’t do: stop at money. It maps the full reward landscape — status, autonomy, identity, fear of sanction, approval — because the financial incentive is often not the dominant lever, and reading only the paycheck misses the structure actually steering the room.

How it works

In colonial Delhi, the British administration had a cobra problem: too many venomous snakes in the city. So they reached for an obvious fix and put a bounty on them — bring in a dead cobra, collect a cash reward. It worked, at first. Snakes came in, money went out, the count seemed to fall. And then enterprising residents noticed something the administrators hadn’t: a dead cobra was now worth money, and cobras are not hard to breed. Quietly, people started farming them — raising cobras for the express purpose of killing them and collecting the bounty.

When the government caught on and scrapped the program, the cobra farmers were left holding stock that had suddenly become worthless. So they did the rational thing and released their snakes. The wild cobra population ended up higher than before the bounty began. The policy had performed flawlessly — it just performed toward the reward it actually offered (“produce dead cobras to be paid for”) rather than the goal the designers had in mind (“fewer cobras in the city”). The two had quietly come apart, and the incentive followed the money every step of the way.

This is the whole lens, and the investor Charlie Munger put it as bluntly as anyone: “Show me the incentive and I’ll show you the outcome.” People respond to the reward they actually face, not the one designers intended — and the gap between the two is where most organizational dysfunction lives. A hospital rated on waiting times finds ways to make the waiting room technically empty. A sales force paid on volume discounts everything to nothing. A team told that quality matters, but promoted on shipping speed, ships fast and patches later. In each case the people are not the problem; they are rational actors optimizing the structure they’re standing in. The actor-character story — they’re lazy, they’re careless, they don’t care — is almost always a placeholder for a reward structure nobody has bothered to map.

The deep move, then, is to treat the structure as the variable rather than the person. Once you map what each actor is genuinely rewarded for — and crucially, that means more than money; status, autonomy, identity, and the avoidance of punishment are all consequential — the behavior stops being a mystery and becomes a design problem. And design problems can be fixed: arrange things so that the easiest path for each actor is also the one that produces the outcome you want. The one discipline that separates this from naïve optimism is to ask, before you change anything, how will people game the new rule? Because they will. New incentives breed new strategies, and the second-order response — the cobra farm that springs up around your bounty — is often the one that decides whether the fix works or backfires.

Framework & implementation

This section uses Ora’s own terms for the parts of an analysis, so that if you open the actual mode and lens files they line up. Each is glossed in plain language on first use.

Pipeline execution

Incentives is one of the always-loaded mental models in the Stakeholder Mapping analysis — it sits in the mode’s ANALYTICAL PERSPECTIVES block under “always loaded,” available as a reasoning tool throughout. It is not the mode’s structure (the catalog lens that supplies the salience grids is stakeholder-analysis-frameworks); incentives is the discipline that keeps each party’s stake honest. The analysis runs at Gear 4, Ora’s most thorough setting — a Depth analyst and a Breadth analyst map the parties in parallel, critique each other, and revise.

Where the lens engages. It activates on its Detection Signals — a system producing outcomes its designers can’t explain; people visibly gaming a metric; a proposed fix that is hiring or training rather than restructuring rewards. Its Application Steps run inside the map: for each party, map the full incentive landscape (money, status, approval, autonomy, avoidance of punishment), distinguish stated incentives from real ones (what the system actually rewards and consequences), and identify where the rewarded behavior diverges from the desired outcome.

What it contributes to the map. Stakeholder Mapping’s job is to inventory the parties, plot them on the Bryson power-interest grid and the Mitchell-Agle-Wood salience classes (power × legitimacy × urgency), and name each party’s stake. Incentives does its sharpest work in the stake section — the mode’s CQ2 demands stakes named at the level of concrete interests (what a party wants and could lose), not role-labels, and that is exactly the incentive question: not “the regulator” but “the regulator whose staff are judged on visible enforcement.” A role is a placeholder; an incentive predicts behavior. It also informs the relationships among parties — alliances and oppositions usually track shared or clashing rewards — and helps surface absent parties whose interests no one is currently paid to represent.

Cross-adversarial evaluation. At Gear 4 each analyst’s reading is critiqued by the other, where the lens’s own failure modes are caught — keyed to its Critical Questions: diagnosing dysfunction as a talent or character problem when it’s structural (people-blame substitution); mapping only money (financial-only mapping); proposing a new metric without modeling the gaming response (gaming-blind redesign); and substituting exhortation for structural change (exhortation substitution). The evaluator presses each: is this stake a real consequence the party faces, or a stated value with no reward attached?

Honesty discipline. The mode keeps confidence calibrated and resists the tidy-diagram bias — dropping low-power-but-legitimate parties for a cleaner picture. Incentives reinforces this: the party with no organized constituency is often precisely the one whose interests no incentive currently protects, and a map that quietly mirrors existing power (the mode’s silent-power-mirroring failure) is one that has stopped asking who is rewarded to speak.

What the analysis will not do. It will not reduce incentives to money, and it will not treat words — a values statement, a training, a culture deck — as if they moved behavior on their own. Where a redesign is proposed, it projects the second-order gaming response rather than assuming actors won’t adapt; an incentive analysis that assumes no one games the rule is not finished.

Origin and evidence

The principle is ancient, but its modern formulation runs through several sources the lens draws on. Charlie Munger’s address “The Psychology of Human Misjudgment” (1995, collected in Poor Charlie’s Almanack) put the “incentive-superresponse tendency” at the top of his catalog of human misjudgment and supplied the canonical “show me the incentive” formulation. Steven Kerr’s 1975 classic “On the Folly of Rewarding A, While Hoping for B” cataloged the recurring organizational pattern of reward systems that pay for one thing while management hopes for another — the academic backbone of the lens. Gary Becker’s The Economic Approach to Human Behavior (1976) extended incentive analysis systematically to non-market behavior (crime, family, discrimination), arguing that people respond to costs and benefits far beyond the marketplace. Levitt and Dubner’s Freakonomics (2005) popularized the case-based version for a general audience. The principle’s close cousins — Goodhart’s Law (when a measure becomes a target, it stops being a good measure) and the principal-agent problem (the formal frame for misaligned authority and execution) — are the structural statements of the same insight.

Applications and common uses

Incentives is a working tool wherever behavior puzzles its designers, used to diagnose why a system misbehaves and to redesign it so the easy path is the right one.

  • Organizational diagnosis. The first question to ask about any persistent dysfunction — turnover, gaming, quality collapse — is what the structure actually rewards; the answer usually dissolves the “people problem.”
  • Policy design. From the cobra bounty to “teaching to the test” to emissions loopholes, public policy is a graveyard of incentives that worked perfectly toward the wrong target; mapping the gaming response in advance is the discipline that prevents it.
  • Metric and compensation design. Every KPI is an incentive, and Goodhart’s Law guarantees that any measure pushed hard enough becomes a target people optimize instead of the goal — so metrics must be designed against their own gaming.
  • Root-cause analysis. When an investigation keeps landing on “human error,” the incentive lens reframes it: what made the erroneous path the rewarded or path-of-least-resistance one?
  • Behavioral and product design. Aligning what’s easy with what’s intended — defaults, rewards, frictions — is incentive design, with the standing caution that extrinsic rewards can crowd out intrinsic motivation in creative or prosocial contexts.

In every case the move is the same: map what each actor is really rewarded for, find where that diverges from the goal, change the structure so the divergence closes, and model how people will game the new structure before you ship it.

Failure modes and when not to use it

The lens’s characteristic ways of going wrong are catalogued in its Common Failure Modes:

  • People-blame substitution. Diagnosing an incentive problem as a talent or character problem. The tell is a proposed fix that’s firing or hiring rather than restructuring rewards. Re-run the diagnosis explicitly mapping what each actor is rewarded for.
  • Financial-only mapping. Treating money as the only consequential incentive. The tell is an analysis silent on status, autonomy, fear, and identity. Map the full reward landscape; the financial lever often turns out not to be the dominant one.
  • Gaming-blind redesign. Proposing a new metric without modeling how actors will optimize against it. The tell is a redesign that assumes people won’t adapt. Project the gaming response and design for it.
  • Exhortation substitution. Using training, culture work, or values statements in place of structural change. The tell is that the intervention is words while the reward structure is untouched. Change the structure; words follow, they don’t lead.
  • Crowding-out blindness. Adding financial incentives where they backfire — creative work, prosocial behavior — and watching performance drop. Distinguish contexts where extrinsic rewards complement intrinsic motivation from those where they corrode it.

When not to reach for it. When the actors genuinely aren’t free to respond — under coercion, extreme constraint, or with no real choices — behavior isn’t tracking incentives and the lens over-explains. When the reward structure can’t be observed or even reasonably inferred, an incentive story becomes unfalsifiable just-so storytelling; say so rather than inventing one. And when the behavior really is about competence or capacity rather than motivation, the incentive frame misdiagnoses a training or resourcing problem as a rewards problem — the one case where “it’s a people problem” is actually true.

  • Stakeholder Mapping — the analysis this lens informs; inventories the parties, plots their power and salience, and names each one’s concrete stake.
  • Principal-Agent Problem — the formal frame for the misalignment incentives diagnoses: an agent rewarded for something other than what the principal wants.
  • Signaling — the flip side: how parties use costly actions to communicate within an incentive structure, and how that shapes what the rewards actually produce.
  • Goodhart’s Law — the sharpest corollary: when a measure becomes a target, people optimize the measure instead of the goal it was meant to track.