THE HUMAN INDEX
BUSINESS Prefrontal cortex · The map of the mind

Planning fallacy: why your tasks always take longer

We know our past projects ran late, yet we expect the next one to go smoothly. Here is what the research shows and one simple question that helps.

No. 003 · 16 November 2026 · 5 min read

2 sources
In this entry

You tell a colleague the report will take two hours. By the end of the afternoon you are still on page three, and the email you meant to send first has not been touched.

The odd part is that this is not new. Last month's report also ran long, and so did the one before. Yet the new estimate felt sensible when you made it.

This habit has a name: the planning fallacy. It is the tendency to expect your own project to go as planned, even when you know that similar projects have usually run late.

What is the planning fallacy?

The term comes from Daniel Kahneman and Amos Tversky, who used it in 1979 1. The classic experimental test came in 1994, from psychologists Roger Buehler, Dale Griffin and Michael Ross 1.

They described the puzzle plainly. People can hold two beliefs at once: that their earlier predictions were too optimistic, and that this time their forecast is realistic 1. In a classroom survey they ran, students said they had finished about two thirds of their previous projects (an average of 68%) later than they expected 1.

33.9Average days students predicted for their thesis (they actually took 55.5)
29.7%Share of students who finished within their own best-estimate time
60.0%Finished on time when asked to link past experience to the task (control: 29.3%)

What did the thesis study find?

The first study used 37 psychology students in the final semester of an honours thesis course at the University of Waterloo in Canada 1. A researcher phoned each student and asked when they would submit. The course coordinator recorded the real date 1.

When asked for their best estimate, students predicted they would finish in 33.9 days on average. They actually took 55.5 days 1. Only 29.7% finished in the time they had predicted 1. Four more students had not finished when the researchers stopped keeping records, two semesters later, so the true gap was probably larger 1.

The students were also asked for a worst case, "if everything went as poorly as it possibly could". Their average answer was 48.6 days, still shorter than the 55.5 days they really took. Fewer than half (48.7%) finished by that date 1.

The forecasts were not useless. Students who predicted longer times did tend to take longer, with a correlation of .77 1. The estimates ranked jobs sensibly but were too short across the board.

The same pattern appeared in everyday tasks. Among academic projects that students completed within a week-long follow-up, the average prediction was 5.8 days and the average real time was 10.7 days 1. Telling students that the researchers would check their accuracy made no difference 1.

Why do we fall for it?

Kahneman and Tversky separated two ways to forecast. The "inside view" works out how this particular task will unfold. The "outside view" asks how long similar tasks took 1.

People mostly use the inside view. When students thought aloud while predicting, about three quarters of their thoughts (0.74) were about future scenarios, against 0.07 about relevant past experience 1. In a later study, 93.5% of students considered plans for the task, but only 9.8% mentioned anything that might get in the way 1.

When past delays did come to mind, students tended to explain them as external, temporary and specific to that occasion 1. If every delay was an exception, the track record never seems to apply to this time.

Outsiders do not make this mistake as much. Observers asked to predict someone else's completion time gave an average of 8.5 days, against 5.5 days from the people doing the task 1. A review for the UK Department for Transport adds a simple point. A plan lists what you intend to do, so setbacks are not part of it and tend to be left out 2.

The effect is not limited to students. The same review cites a meta-analysis of 258 transport projects worldwide in which actual costs were on average 28 percent higher than estimated 2.

What are the limits of the research?

The original studies used small groups of university students, and the thesis result rests on 37 people 1. The wider evidence is larger, but the review notes that it is mostly about cost and time overruns in project work 2.

The fixes are less well tested than the bias itself. In one study, students who recalled past assignments and then linked them to the new task predicted later finish times. 60.0% of them finished within their predicted time, against 29.3% of the control group 1. But the authors found these predictions were less biased, not more accurate for each individual 1. The government review says very little has been empirically tested, so most suggestions are well-founded ideas rather than proven methods 2.

One approach does have support in project work. Reference class forecasting compares your project with the spread of outcomes from similar past projects, and the review reports it is more accurate than conventional forecasting 2. It also lists drawbacks, including the need for data on many past projects 2.

A plan tells you what you intend; your track record tells you what tends to happen.

How to use the planning fallacy

  • Start from your own record. Before you promise a date, find the last two or three similar tasks and note how long they really took. Begin from that, not from the best-case story.
  • Link the past to the present. Do not just remember that earlier work ran late. Ask what, specifically, would make this one different, and whether you have evidence for it.
  • Ask someone else. People judge other people's tasks more cautiously than their own. Have a colleague or friend give an independent estimate before you commit.
  • Write down what could go wrong. List the obstacles as carefully as you list the steps. This is the part our natural planning leaves out.

None of this guarantees an accurate estimate. It makes it harder to confuse a hopeful plan with a forecast.

Sources

  1. Exploring the "Planning Fallacy": Why People Underestimate Their Task Completion Times. Journal of Personality and Social Psychology (Buehler, Griffin and Ross, 1994) Read it
  2. A review of optimism bias, planning fallacy, sunk cost bias and groupthink in project delivery and organisational decision making. Behavioural Insights Team for the UK Department for Transport Read it

Get the week's best ideas in your inbox.