A risky decision works out, and the person who made it gets called brilliant. A carefully considered decision goes badly, and the same room suddenly knows what should have been done.
The result has changed how the decision looks. That is understandable. It can also leave a business learning the wrong lesson from a very good week.
Results matter. A business cannot pay its people with a beautifully documented decision process. But if we want to improve the next choice, we need to separate what happened from what could reasonably have been known before it happened. Otherwise, success can make a poor habit harder to question.
The rescue can disappear inside the success
Imagine a company launching a campaign before checking whether its operations team can handle the response. The creative work connects. Orders arrive. People stay late, postpone other work and improvise their way through delivery. The sales report looks excellent.
At the review meeting, someone says the campaign proves the value of moving quickly. The next launch should be even more ambitious.
This is an illustrative scenario, not a report of a client experience. Its purpose is to make the learning problem visible. The sales result may be real, and the creative work may deserve praise. Yet the result alone cannot tell us whether skipping the capacity check was sensible. The team may have succeeded despite that choice.
If the review credits the shortcut and overlooks the rescue, the company can build next year’s plan around an approach its staff cannot sustainably repeat. A win has become evidence for the wrong behaviour.
Outcome bias changes the verdict
In their 1988 paper, Jonathan Baron and John Hershey examined outcome bias in decision evaluation. Across five studies, undergraduate participants evaluated descriptions of medical decisions or monetary gambles. They had the relevant information available to the original decision maker. Knowing that a decision ended favourably still led them to evaluate its thinking or the decision maker more favourably.
The study does not establish what is happening in any particular management team. Its participants and scenarios were different. It does give us a reason to be careful when the final result starts doing all the explaining.
Outcome bias is related to hindsight bias, but they are different mistakes. Hindsight bias makes an event seem more predictable after we know what happened. Outcome bias changes our evaluation of the decision because we know how it turned out. A review can contain both: the winner was obviously right, and the loser should obviously have known better.
My discussion of hindsight bias in The Divided Brain concerns that first problem, the way later knowledge can distort our view of earlier judgment. Outcome bias adds a separate, externally researched question: are we grading the quality of the choice, or simply reacting to its ending?
Review the choice before the ending
A practical response starts before the decision is made. Keep a short record of what you expect, why the choice seems worthwhile and what could go wrong. You do not need a new committee or a complicated scoring system. A few honest sentences can preserve something memory may later smooth over.
For the campaign, that record might say: we expect demand to rise, we have enough capacity within an agreed range, and if orders exceed it we will adjust delivery promises before accepting more work. Those are prospective judgments. They can be compared with what actually happens.
When the review comes, start with that earlier record. Consider whether the team used the information it had, took reasonable precautions and understood the trade-off. Only then bring in the ending and ask what it teaches.
That order matters because knowing the result is difficult to undo. A written record does not make anyone immune to bias. It gives the conversation something firmer than an account reconstructed after everyone has seen the sales chart.
Luck is not a reason to stop learning
Separating decision quality from outcome should not become a convenient escape hatch. Someone cannot label every win skill and every loss bad luck. Nor should a team ignore a pattern of disappointing results because its process sounds impressive.
Repeated outcomes can reveal that an assumption was wrong. An apparent one-off may expose a risk that deserves attention. The task is to examine those possibilities rather than choose whichever explanation protects the person presenting the report.
A useful review can hold two thoughts together: the decision was reasonable with the information available, and the information or method needs to improve. Or the result was excellent, and one part of the approach should not be repeated.
For the imagined campaign, the right response might be to keep the creative idea, recognize the people who delivered it and require a capacity check next time. That is a more useful lesson than either declaring everything brilliant or dismissing the whole success as luck.
What leaders reward becomes the next experiment
The language used after a result shapes what people feel encouraged to do next. If only the headline number gets attention, people may learn to conceal the strain and uncertainty behind it. If every unsuccessful attempt attracts blame, they may avoid worthwhile experiments whose outcomes cannot be guaranteed.
A leader can praise the result while also asking what was repeatable, what was rescued and what depended on circumstances outside the team’s control. Recognizing those differences gives people a better account of what good work looks like. It also makes the next review less dependent on whoever tells the most flattering story.
Success should give a business room to examine its decisions, not a reason to exempt them from examination.
Celebrate the win. Before turning it into a rule, check what you are about to learn from it.
Research source: [Baron and Hershey, Outcome bias in decision evaluation (1988)](https://pubmed.ncbi.nlm.nih.gov/3367280/).
Paul Larche is the author of The Divided Brain (CanReads 2026 Business/Marketing Winner | BookLife Prize 2025 Semifinalist, 10/10 Score, Editor’s Pick).