Ask anyone how long something will take, then quietly double it. You will be closer to the truth than they were, and they will be mildly offended that you were. This is not pessimism. It is arithmetic learned the hard way, and almost nobody applies it to themselves.
The reason has a name. The planning fallacy is our systematic tendency to underestimate the time, cost and risk of our own projects, even when we have lived through the exact same delay before. The renovation that was six weeks becomes six months. The report due Friday is filed at midnight on Sunday. The startup that needed eighteen months of runway is, at month nineteen, drafting an apologetic email to investors. We do not learn, because each time we are certain that this time is different. It rarely is.
You meet the fallacy in its purest form at the counter. It is the barista who promises ‘two minutes’ for a flat white that has never, not once, taken two minutes. There is a queue. There is steam. There is a milk jug being rinsed and a customer ahead asking whether the oat is the new oat or the old oat. Two minutes is not a measurement; it is a wish spoken with a confident face. And the remarkable thing is that you believe it, every single morning, despite a lifetime of standing there at minute four watching the wish evaporate.
What is actually happening is a quiet swap of evidence. When we estimate our own brew, our own deadline, our own grand plan, we build the forecast from the best-case story playing in our heads: the smooth pour, the green lights, the morning where nothing goes wrong. The APA defines the planning fallacy as exactly this kind of unrealistically optimistic forecast, a subspecies of the broader optimism bias that nudges us to overestimate the good and discount the bad. We plan from the vivid inside story rather than the dull, reliable base rate of how similar things have actually gone. The base rate is boring. The base rate is also correct.
Daniel Kahneman and Amos Tversky, who first named the effect, spotted the obvious tell. Outsiders are better at estimating your timeline than you are, precisely because they are too lazy to imagine your special circumstances. They just ask how long these things usually take. Kahneman later turned this into a working cure with the economist Dan Lovallo, arguing in a now-famous Harvard Business Review piece on optimistic forecasts that executives torpedo billion-dollar projects by taking the ‘inside view’ of their own brilliance instead of the ‘outside view’ of what comparable projects actually delivered. The fix, sometimes called reference-class forecasting, is almost insultingly simple: stop asking how long your project will take, and start asking how long projects like yours took. Then trust that number, even though it stings.
It stings because it feels like betrayal. The optimistic estimate is not just a number, it is a small act of faith in your own competence, and the base rate quietly informs you that your competence is statistically unremarkable. So we keep choosing the wish. We say two minutes. We say six weeks. We say the launch is on track, right up until the launch teaches us, again, that tracks have curves. The cost is not only late coffee. It is budgets blown, trust spent, and the slow erosion of the one currency a promise is supposed to buy: belief that the next promise means anything.
The escape is not to become gloomy. Gloom is just optimism wearing a different costume, equally untethered from the data. The escape is to widen the sample. Look at the last ten brews, the last ten deadlines, the last ten times you said ‘quick’. Let the dull average overrule the shining exception. The story in your head is a sample size of one, and it is auditioning hard for the role of typical case.
Optimism is lovely, and the clock does not care. The honest estimate, the one that earns groans in the meeting, is almost always the one you were embarrassed to say out loud. Say it anyway. The brew will still take longer than that, but at least, for once, you will not be surprised.
