Good Decisions, Bad Outcomes

Key takeaways

• A good decision and a good outcome are different things. Ecommerce, like poker, pays out on a blend of skill and luck, and the two come apart more often than we admit.
• Outcome bias leads us to praise lucky gambles and punish sound calls that happened to lose, which trains people to optimise for looking right rather than for deciding well.
• Judging a decision means asking what was known when it was made, not what you happen to know now.
• Post-mortems work best when they revisit the original reasoning, not only the result, which is exactly why most organisations avoid them.
• The riskiest moment is often the one just after you were right, because success by luck teaches a confident, wrong lesson.

A good decision and a good outcome are different things

Most of us judge a decision by how it turned out. It is the natural move, and it is a trap. Ecommerce is closer to poker than to chess: you can play a hand well and lose it, and play a hand badly and win. A founder who bets the whole quarter on one influencer launch and gets lucky made a poor decision that happened to pay. A founder who diversified sensibly and got caught by a freak ad-platform outage made a good decision that happened to lose. Score only the result and you cannot tell those two people apart, and you will end up quietly admiring the gambler.

Outcome bias trains people to look right rather than decide well

Kahneman called this outcome bias: we blame decision-makers for sound choices that worked out badly, and hand too much credit to reckless ones that came good. The organisational cost is quiet and large. People learn what gets rewarded, and if the reward goes to being right rather than to reasoning well, they will start managing how things look rather than how decisions are made. They will dodge sensible bets that carry visible downside and chase fragile ones that usually work, because "usually works" keeps the outcomes looking good, right up until the day it does not.

Judge the call on what was knowable at the time

The fairer question is the harder one: given what could be known when the decision was made, was this a reasonable thing to do? The narrative fallacy from the Your Brain posts is the enemy here, because hindsight rewrites the odds. Once you know the launch flopped, every warning sign looks obvious and the people who missed them look foolish. They were not necessarily foolish. They were working without the answer key you now hold. A discipline I have borrowed from people who take this seriously is to ask, after a loss, "did I do my best with what I had?" rather than "did it work?" The first keeps you honest. The second just blames the weather.

Post-mortems should revisit the reasoning, not just the result

Charlie Munger admired Johnson & Johnson for a habit most firms cannot stomach. When an acquisition went wrong, they made everyone go back to the original presentations, the ones that had argued for the deal, and read them again. Most companies do the reverse: a disaster happens and the paperwork that caused it quietly vanishes, because nobody wants to be seen near it. That is close to the worst possible response, because the one thing that improves decisions is examining the decision, not just the wreckage. The reasoning is where the lesson lives.

I watched a team do this well once, after a discount strategy that lost money. Instead of hunting for someone to blame, they pulled up the original logic and found it had been sound on the numbers available at the time, undone by a competitor move nobody could have seen. They kept the approach, adjusted for the new risk, and made the money back the next year. Had they scored it on the outcome alone, they would have thrown away a good decision for the crime of one bad night.

Being right is the most dangerous outcome

Peter Bernstein had a line I keep close: the riskiest moment is when you are right, because that is when you are most tempted to overstay a good decision and to mistake luck for skill. Nothing blinds quite like success you did not have to work for. The founder whose gut call paid off learns to trust the gut, which serves them well until the gut walks them off a cliff.

Which leaves a question this post cannot resolve, and which the organisation tier of the series takes up: how do you build a company that rewards good decisions even when they lose, in a world where investors, staff, and your own nervous system all reward good outcomes? I have seen it managed in small pockets and for short stretches. Keeping it up when the results are bad and everyone wants a head on a spike is the part I still find hard.