Outcome Bias: Why Judging Decisions by Results Teaches You to Be Lucky, Not Skilled

One of your people ran a customer call last week and it went badly. The account is unhappy, your boss has heard about it, and you’ve already formed a view: that was a bad piece of work. When you replay it in your head, the things she did wrong are easy to find. She led with the new pricing too early. She didn’t loop in the account manager. She pushed for a decision when she should have listened.

Now slow down and ask a harder question. If that same call had landed well, if the customer had said yes and thanked her for being direct, would you be cataloguing those same moves as mistakes? Leading with pricing would read as confidence. Skipping the account manager would read as ownership. Pushing for a decision would read as closing instinct. Same person, same preparation, same choices, same information she had going in. The only thing that changed is how the customer happened to react, which is the one part of the call she didn’t control. And yet that part is doing almost all of the work in your judgement of her.

The evidence

This is outcome bias, and it has been pinned down cleanly. Give people the exact same decision, described in the exact same words, and change only how it turned out. Those who are told it ended well rate the decision-maker as more competent, more careful, and more sensible than those told it ended badly. The reasoning on the page is identical. The result is the only thing that moved, and it moves the verdict.

Here is the part that should worry you, because it’s the part you’d expect your own intelligence to protect you from. The effect holds even when people are told outright that the outcome came down to luck, to factors the decider couldn’t have known or influenced. You can hand someone the information that the bad result was nobody’s fault, and they still mark the person down for it. Knowing luck was involved doesn’t switch the bias off. It barely dents it. The outcome writes itself over your assessment of the process before the conscious, fair-minded part of you gets a vote.

There’s a second force stacked on top, and it explains why the call that went wrong grips you harder than any call that went right. A loss lands heavier than an equivalent gain. So a bad outcome makes the decision look mistaken and then makes it feel reckless, careless, the kind of thing a serious person wouldn’t have done. The sting recruits the story to match it. By the time you sit down to give feedback, you’re prosecuting the outcome and calling it her judgement.

How it works

The engine underneath all of this is how you build stories from events. Your mind takes a result and reaches backward for the causes, and once it has an ending, it sorts everything before it into the things that led there. A good ending turns the preceding choices into smart moves. A bad ending turns those same choices into the warning signs you should have seen. The sorting feels like understanding. It feels like you’ve worked out what went wrong. But you’ve imposed a pattern that the ending chose for you, and a different ending would have handed you a different pattern from the identical facts.

The trouble is that the outcome blends three ingredients you care about keeping separate: the quality of the reasoning, the luck of the draw, and the situation she was handed. A great process in a hostile situation can still lose, while a weak process handed an easy one can win. If you read straight off the result, you can’t tell which you’re looking at, so you end up rewarding and punishing the luck and the context as if they were the skill.

When you judge a call by how it landed, your team learns to avoid any sound decision whose downside you’ll be able to see.

That last point is where this stops being a private thinking error and starts costing you a team. When the people around you learn that a bad outcome gets punished no matter how sound the call was, they draw the rational conclusion. They stop making the calls where the downside is visible, even the calls with strong odds in their favour, because a 30% chance of a bad result that you’ll judge them for is a 30% chance of looking incompetent. You wanted bold, well-reasoned bets, and what you trained for was safe, defensible inaction.

How to use it

Start by separating what she controlled from what she didn’t, before you say a word about the call. Take the failed customer conversation and split it in two. In one column, the process: what she knew going in, the options she weighed, the reasoning behind leading with pricing, why she made the choices she made. In the other, everything outside her hands: that the customer had just been burned by a competitor, that procurement had quietly frozen spending that week, that her main contact was having a terrible day. Judge her only on the first column. The second one is real, it mattered, and it isn’t hers to answer for.

The test that keeps you honest is the flip. Before you deliver any verdict, ask what you’d be saying if the result had gone the other way. If a good outcome would have earned the same choices a “well played,” then a bad outcome doesn’t earn them a “what were you thinking.” The choices didn’t change. If you’d praise it when it works and slate it when it fails, you’re scoring the dice rather than the play she actually made. Run that flip out loud in your own head before the conversation, because the bias is fast and your fairness is slow, and the flip is how you give the slow part time to arrive.

There’s a harder case the split is built for, and it’s the one most people get backwards. Separating process from outcome is not a way to wave off every failure as bad luck. Sometimes the process really was sloppy, and the bad result is the honest signal. The way you tell the difference is to ask the question at the decision point, not the outcome point: given only what she knew at the time, with the information she actually had in front of her, was the reasoning sound? If it was, a bad result is luck and your feedback should protect the reasoning. If the reasoning was thin even on the day’s information, then the process failed, the result is fair evidence, and the split doesn’t rescue it. The exercise is the same in both cases and the verdict comes out opposite, and the only thing that decides which way it falls is the reasoning at the time.

Why it matters

Most workplaces run on outcomes, because outcomes are easy to see and reasoning isn’t. Results get the bonus and the promotion, and a failure gets the cold shoulder however careful the thinking behind it was. The signal that sends is unmistakable, and your people are good at reading signals. They learn to optimise for what gets rewarded and to dodge anything whose downside you’d be able to hold against them. You get caution dressed as prudence and a team that has stopped taking the bets you hired them to take.

Pulling the result apart from the call is, in that setting, the most useful thing you can do as the person whose verdict everyone is watching. It’s how you stop letting a bad week of luck masquerade as evidence about someone’s judgement, and it’s how you keep someone’s one good fluke from convincing you they can walk on water. Do it consistently and the people around you learn the one thing that actually makes them braver and better: that you’ll back a good call even when it loses, and you’ll question a lucky one even when it wins. That is a rarer thing to work for than it sounds, and it changes how people decide when you’re not in the room.