You Have Been Wrong About How Long That Will Take, and You Will Probably Be Wrong Again

The report that was supposed to take an afternoon eats the whole week. The renovation estimated at two months is still ongoing at month five. What makes this pattern strange is not that it happens once. It is that it keeps happening to the same people, on similar tasks, even when they can clearly remember the last time they were wrong about the exact same kind of estimate.

Daniel Kahneman and Amos Tversky named this the planning fallacy in 1979: the tendency to predict a task will go roughly as planned, even while holding a much more pessimistic, and more accurate, view of how similar tasks have gone in the past. A frequently cited 1994 study asked psychology students to estimate how long their senior thesis would take. The average estimate was under 34 days. Only 30 percent actually finished within their own predicted timeframe. These were not naive predictions from people without relevant experience. Most had watched previous cohorts run late on the exact same kind of project.

Kahneman and later collaborator Dan Lovallo traced this to a specific cognitive habit they called the inside view: when estimating a task, people naturally focus on the specifics of this particular project, this time, this plan, largely ignoring the outside view, the broader pattern of how this entire category of task has actually gone historically, for you and for other people. Research on this bias has found something that makes it more stubborn than it looks: people can accurately state that similar projects usually run long, while still confidently predicting this one will be different, because the inside view of the current task feels more vivid and more informative than the abstract outside pattern, even when the outside pattern is the better predictor.

Part of what sustains this pattern is what researchers call presentism: predictions get made in a calm, focused state, imagining a best-case version of the work ahead, which systematically underweights the interruptions, unexpected complications, and ordinary friction that showed up every previous time and will very likely show up again. Research extending the original concept, sometimes labeled the planning fallacy writ large, has found the same basic bias applies well beyond time estimates too, extending to underestimated costs and overestimated benefits across everything from home renovations to major public infrastructure projects, where cost overruns following optimistic initial projections are a well documented, recurring pattern across decades of data.

The fix Kahneman and Lovallo proposed is specific and somewhat unnatural to actually do: deliberately adopt the outside view before finalizing an estimate. Instead of reasoning from scratch about how long this particular task should take, look at how long this category of task has actually taken before, for you specifically if you have the data, and adjust your estimate toward that number rather than your gut feeling about this one. It will almost always mean a longer number than feels right in the moment. Given how consistently the research finds people underestimate in the other direction, that discomfort is itself a fairly reliable sign the outside view estimate is closer to correct than the inside view one.