Outside-View Base-Rate Planning
Forecast projects from comparable outcomes before trusting an inside estimate
- Difficulty
- Moderate
- Time to result
- ~days to results
- Steps
- 6
- Confidence
- 97%
Outside-View Base-Rate Planning counters the planning fallacy by beginning with comparable projects rather than the current team's confidence. The planner defines an appropriate reference class, gathers actual durations and abandonment rates, and uses those results as the initial forecast. Only then should the estimate be adjusted for specific differences in scope, staffing, technology, or constraints. This reverses the usual process in which teams focus narrowly on their own competence, motivation, and ideal sequence of work. The framework also includes failed and unfinished projects, preventing survivor bias from making the baseline too optimistic. Its output is a range with a probability of completion, explicit contingency, and documented assumptions—not a single confident date detached from historical evidence.
Origin
Extracted from Masters in Business through Jason Zweig's account of Daniel Kahneman's textbook project, which was forecast at roughly a year but took many years despite explicit awareness of the planning fallacy.
Core principles
- 01Comparable outcomes are stronger anchors than enthusiasm
- 02The inside view systematically narrows attention
- 03Completion rates matter as much as average duration
- 04Plans need ranges and failure probabilities
- 05Expert awareness does not eliminate planning bias
How to run it
- 1
Define the reference class
Describe the project in terms that allow comparison with prior efforts of similar scope, novelty, and organizational complexity.
Pro tip Use several plausible reference classes and test how sensitive the estimate is to each.
Watch out A class that is too broad or too flattering will produce a misleading baseline.
- 2
Collect completed and failed cases
Gather actual durations, costs, scope changes, and abandonment rates from comparable projects.
Pro tip Include unfinished work rather than studying survivors alone.
Watch out Ignoring abandoned projects systematically understates risk.
- 3
Calculate the base rate
Determine the median duration, realistic range, and probability of non-completion.
Pro tip Prefer distributions and percentiles to a simple average.
Watch out Do not discard inconvenient cases without a documented reason.
- 4
Anchor on the outside view
Use the historical distribution as the initial forecast before discussing the current team's special qualities.
Pro tip Present the base rate before inviting estimates from project advocates.
Watch out Starting with an optimistic inside estimate makes later adjustments too small.
- 5
Adjust for specific differences
Modify the forecast only where evidence shows the present project differs materially from the reference class.
Pro tip Quantify each adjustment and assign an owner to its assumption.
Watch out Claims that the team is smarter or more motivated are not sufficient evidence.
- 6
Add contingency and review gates
Set schedule and budget reserves, plus milestones where assumptions and the reference class will be revisited.
Pro tip Publish a range and confidence level rather than one ceremonial deadline.
Watch out A contingency that stakeholders quietly consume before work begins is not a contingency.
In the wild
A team initially estimates a database migration at eight weeks. Five comparable migrations took between four and nine months, and one was abandoned. The team anchors on the median, adjusts downward for better tooling, and adds review gates for data-quality risk.
→ Leadership receives a defensible five-to-seven-month range and an explicit failure risk instead of an optimistic promise.
A team estimated that rewriting a psychology textbook might take roughly a year. Comparable teams reportedly averaged about four years among those that finished, while many never completed the work. The project ultimately took substantially longer than its initial estimate.
→ The story demonstrates how the inside view can overpower relevant historical evidence.
Common mistakes
Choosing flattering comparisons
Teams often select unusually successful projects as analogues. Use the full relevant distribution.
Excluding non-completions
Omitting abandoned projects hides the true probability that the current effort may also fail.
Assuming awareness cures bias
Kahneman's example shows that understanding the planning fallacy does not guarantee an accurate forecast.
Is it for you?
Best for
It is best for teams estimating complex projects with uncertain requirements and a history of delays.
Not ideal for
It is less useful for routine tasks with stable throughput and abundant direct operational data.
From the transcript
“we don't look at the base rate.”
“the single most important question is what is the base rate?”
“I guess forty or fifty percent”
From the episode
An Interview With Jason Zweig: Masters in Business (Audio)
Jason Zweig