Incentive Super-Response Tendency

Imagine Leena leading customer support for a subscription software company. She wants customers to get useful answers sooner, so she offers the team a monthly reward when its average ticket-resolution time falls below a set threshold.

The dashboard soon turns green. Ambiguous tickets are closed after the first reply. Complicated requests are split into new tickets. Edge cases receive a link to the general help page. No one has to be dishonest for this to happen. The rule has made speed visible, countable, and worth pursuing.

Two weeks later, reopened tickets and repeat contacts are rising. The score improved. Durable resolution did not.

That gap is the problem described by Incentive Super-Response Tendency.

What is Incentive Super-Response Tendency?

Hello to Halo treats Incentive Super-Response Tendency as a practical name for a specific failure mode: once a benefit depends on a score, behavior may converge on the score even when the result it stands for does not improve. Rolf Dobelli's publisher notes trace the label to Charlie Munger's writing on the power of incentives. It is better understood as a mental-model warning than as a standardized clinical diagnosis.

The underlying incentive-design problem has a stronger research base. In work on multitask jobs, economists Bengt Holmstrom and Paul Milgrom showed why rewarding a measurable activity can move attention away from work that is harder to measure. Their model also accounts for performance measures that can be influenced by actions the organization did not intend to reward.

The useful question is therefore narrower than “Are incentives good or bad?” It is: What behavior does this particular rule make worthwhile?

How a metric pulls effort away from the goal

An incentive system contains at least two things:

• the outcome someone actually wants
• the evidence used to decide whether that outcome happened

Those two things need not be identical. A support manager wants a customer's problem solved. A dashboard can record when a ticket closes. Closure time is useful information, but it does not capture whether the answer was accurate, whether the problem returned, or whether a difficult case was pushed elsewhere.

When the reward depends heavily on that partial measure, the measure gains practical weight. People may spend more effort on the counted task and less on uncounted quality. They may also find a shortcut that improves the number without improving the result. Sometimes that is deliberate gaming. Sometimes it is ordinary prioritization. Sometimes it is simply following the written rule.

This is why stronger incentives are not automatically better incentives. The risk depends on what the measure leaves out, how easily it can be improved without real progress, and what competing work loses attention.

Incentives can work when the measure fits the work

The warning should not be stretched into “rewards always backfire.” They do not.

Edward Lazear's study of a vehicle-glass company found that output and profits rose after the company adopted piece-rate pay. Part of the gain came from workers changing their effort, and part came from which workers joined or stayed with the company. In that setting, the rewarded output tracked useful production well enough for the system to improve performance.

The contrast matters. A narrow speed target is risky when speed can be separated from quality. A production measure may work better when units are clearly defined, quality is observable, and workers are not being pulled away from important unmeasured tasks.

Uri Gneezy, Stephan Meier, and Pedro Rey-Biel reached a similarly careful conclusion in their review of behavioral incentives: effects depend on design, the form of the reward, its interaction with social and intrinsic motivations, and what happens after the reward is removed.

Three mistakes to avoid

Treating the metric as the outcome

A number is evidence about performance. It is not the performance itself. Before attaching a reward, write down what the number fails to observe.

Assuming a distorted result proves bad intent

An incentive changes priorities even when everyone acts in good faith. Start by inspecting the rule and the available shortcuts before deciding that the people involved are the problem.

Adding more pressure when the measure is incomplete

A larger reward increases the reason to pursue the measured result. If that result can drift away from the real goal, more pressure can widen the gap. Improve the measurement design before increasing the stakes.