Estimate timelines based on actual time taken in past similar tasks, not on plans

4. Estimate timelines based on actual time taken in past similar tasks, not on plans

Cost: No cost at all. It only requires a few minutes of review to look up how long similar tasks took previously.

In plain terms: Students who estimated they would finish a paper in 33.9 days actually took 55.5 days to complete it; only 29.7% of them met their own estimates. Another group of students was asked to first recall how long similar tasks had taken them in the past before making their estimates. The proportion of students who finished on time rose from 29.3% to 60.0% in this group.

Benefit: Students originally expected to finish their papers in 33.9 days, but the actual time taken was 55.5 days. Only 29.7% of them completed their work by their own estimates. When researchers asked another group of students to first relate past experiences to the current task before estimating, the proportion of students finishing on time rose from 29.3% to 60.0% compared to the control group. This optimism of underestimating timelines no longer appeared in this group.

Evidence grade: B

Sources: Buehler, R., Griffin, D., & Ross, M. (1994). Exploring the "planning fallacy": Why people underestimate their task completion times. Journal of Personality and Social Psychology, 67(3), 366–381. <https://doi.org/10.1037/0022-3514.67.3.366>; Flyvbjerg, B. (2006). From Nobel Prize to project management: Getting risks right. Project Management Journal, 37(3), 5–15. <https://doi.org/10.1177/875697280603700302>; Halkjelsvik, T., & Jørgensen, M. (2012). From origami to software development: A review of studies on judgment-based predictions of performance time. Psychological Bulletin, 138(2), 238–271. <https://doi.org/10.1037/a0025996>

Notes: Buehler’s study involved only university students in Canada; it was a single study with a small sample size of 37 participants, with roughly 40 participants in each group. Estimating timelines based on historical data from similar projects is known as reference class forecasting; specific engineering applications are described in Flyvbjerg’s work. Halkjelsvik’s review of multiple studies shows that underestimating timelines is more common than overestimating, though the magnitude of the error varies widely depending on the task and research methodology.