CompanyGraph
An incentive is any consequence that changes the value of a choice for the person making it. Pay matters, but so do promotion, deadlines, permissions, attention, and the risks that someone else bears.
Intentions do not operate alone
Organizations state goals such as safety, quality, long-term growth, or customer trust. People then choose actions within budgets, measurement systems, reporting lines, and time limits. An incentive structure is the set of consequences attached to those choices: what earns money or status, what creates delay, what triggers review, and what can be shifted to someone else.
It is too strong to say that incentives always predict behaviour better than intentions. Professional norms, law, identity, and judgment also matter. The useful claim is narrower: when a system repeatedly produces an outcome that conflicts with its stated purpose, inspect the rewards, constraints, and information available to the people making the decisions.
How the structure changes action
A salesperson paid only on bookings may discount heavily, choose easy customers, or pull a future order into the current quarter. A factory manager rewarded on output may run equipment past a maintenance interval. A researcher evaluated on publications may select projects with a higher chance of a positive result. None of these outcomes requires a dishonest person. The measurement, time horizon, and consequence make some actions easier to justify than others.
Incentives operate through several mechanisms:
- Measurement: what is visible receives attention, while unmeasured quality can be neglected.
- Timing: a near-term bonus can outweigh a cost or failure that arrives after the decision-maker leaves.
- Risk allocation: people take more risk when they keep the upside and someone else absorbs the downside.
- Authority: a target without the tools or permission to change the result creates pressure without a feasible remedy.
- Information: the person closest to the work can exploit a metric or report a condition that managers cannot independently observe.
A documented emissions case
The Volkswagen diesel-emissions case shows how a technical target can become an organizational constraint. The U.S. Environmental Protection Agency's Volkswagen enforcement record describes defeat devices that caused vehicles to meet laboratory limits while emitting far more nitrogen oxides in normal driving. The record establishes the regulatory violation and the physical difference between test and road conditions. It does not, by itself, reveal every internal decision or assign the result to one bonus plan. It does show how a performance requirement, development pressure, and the ability to control the test boundary can produce a different outcome from the public objective.
Why simple fixes fail
Adding another metric can help, but it can also create a larger target system to optimize. A quality score paired with a speed score may improve balance; it may also encourage teams to redefine difficult cases or shift work to another department. The design question is not “how many metrics?” but whether the important unmeasured consequences are visible to someone with authority to correct them.
Alignment is also relative. A manager may be rewarded for annual profit, a plant for monthly output, a procurement team for unit price, and a safety group for incident counts. Each can make a defensible local choice that imposes cost or risk on another part of the company. The resulting conflict is structural and can persist even when everyone is trying to follow the stated strategy.
How to inspect an incentive system
- Start with the decision. Which action is the organization trying to influence, and who can actually take it?
- List the consequences. What affects pay, promotion, budget, reputation, approval time, or job security?
- Map the omitted burden. Who bears the customer, safety, maintenance, legal, or long-term cost that the target does not record?
- Check the horizon. When does the reward arrive, and when would a failure become visible?
- Test the feedback. Can frontline information reach a person who can change the target, or does the system punish bad news?
- Look for independent outcomes. Compare the target with cash collection, retention, defect rates, safety, complaints, and physical condition.
Good incentives do not eliminate judgment or trade-offs. They make the desired trade-offs easier to sustain and make the costs of gaming visible. Weak incentives can produce excellent reported metrics and poor operating results; strong intentions cannot reliably repair a structure that rewards their opposite.