Regression to the Mean

Regression to the mean explains why extreme results often move back toward more ordinary levels on the next observation. When an outcome is unusually high or unusually low, it frequently reflects a combination of underlying signal and temporary noise. Because the noise component is unlikely to be equally extreme again, the next result tends to look more typical. 1

Why leaders misread improvement and decline

The most common mistake is to explain every sharp reversal as if management action caused it. A poorly performing branch receives attention, then performs somewhat better and leaders conclude the intervention worked. A star sales team has an extraordinary month, then returns to ordinary levels and leaders conclude motivation slipped. In both cases, part of the movement may simply be regression to the mean.

This matters because extreme results are often unstable by nature. They may contain luck, one-off conditions, unusual timing, or random fluctuation. Once those temporary forces fade, outcomes drift closer to normal even if the system itself did not change very much. The intervention may still matter, but the raw before-and-after contrast exaggerates what really happened.

That is why extreme baselines are so dangerous for judgment.

Where the idea shows up in practice

Regression to the mean appears in investing, medicine, education, operations and talent evaluation. Investors chase funds after spectacular performance only to be disappointed by more ordinary returns. Schools reward teachers after unusually strong cohorts and punish others after unusually weak ones without separating structural quality from statistical swing. Executives often redesign incentives based on single periods that are not representative.

The pattern is especially risky when leaders act on anecdote instead of repeated evidence. A dramatic change feels meaningful because it is emotionally salient. But salience is not the same as causation. If the starting point was extreme, some reversal was likely even without heroic action.

That does not make interventions irrelevant. It makes evaluation harder.

How to tell reversion from real impact

Leaders should ask whether the observed change persists, whether similar cases changed in the same way without the intervention and whether the starting observation was unusually extreme. Repeated measurement matters because true improvement tends to endure more consistently than random snapback. Comparison groups matter because they reveal whether the same movement would have happened anyway.

The goal is not statistical perfection in every managerial decision. It is better causal discipline. A simple pause before assigning credit or blame can prevent many false stories from hardening into policy. The more volatile the system, the more this discipline matters.

Leaders who understand regression to the mean become slower to overreact and faster to seek better evidence.

What leaders should remember

Leaders should be particularly cautious whenever action follows an outlier. That is exactly when natural reversion is most likely to create the illusion of influence. Praising, punishing, restructuring, or investing on the basis of one dramatic swing can entrench weak learning.

The enduring lesson of regression to the mean is simple. Extreme outcomes often move back toward typical levels, so observed improvement after a bad result or decline after a great result is not automatic proof of the intervention that happened in between. 2, 3

• 1Regression toward the mean
• 2Regression fallacy
• 3Thinking, Fast and Slow discussion of regression