The project that got funded is the one whose estimate was furthest from reality. That isn’t bad luck. It’s selection.
Every quarter, a room of executives looks at a stack of proposals and funds a few. The ones that win have the best numbers: the highest return, the lowest cost, the shortest timeline.
Ask yourself what it takes to have the best numbers in a room full of people who all want their project funded. Sometimes it’s the best project. More often it’s the most optimistic estimate.
In Part 1 we looked at how the lowest bid wins procurement because it’s the estimate that missed the most. The same thing happens one level up, where projects compete for budget instead of vendors competing for contracts.
Read the Part 1: Why the Lowest Bid Costs the Most
What is optimism bias?
Daniel Kahneman and Amos Tversky named the planning fallacy in 1979. People estimate a task from the inside, using their own plan, instead of from the outside, using how similar efforts actually turned out. The inside view is seductive because it’s specific: these people, these steps, this timeline. The outside view is dull because it’s statistical.
Optimism bias is the systematic tilt that results. We underestimate cost and duration and overestimate benefits, and we do it consistently, not randomly. That word matters. If estimating errors were random, they would cancel out across a portfolio. They don’t cancel. They stack in the same direction.
Why 73% of Projects Fail: How Burn Rates Changes Everything About Learning Project Management
Why does the funded project end up being the most optimistic one?
Because approval is an auction. Several teams estimate uncertain futures. Their estimates scatter. The most favorable one wins the budget. Being selected is evidence that your estimate was extreme.
Bent Flyvbjerg, who has studied large-project outcomes for decades, calls the result “survival of the unfittest.” The projects that look best on paper get built because they look best on paper, and they look best on paper because their numbers were least connected to reality.
Two forces drive this, and you usually can’t tell them apart from the outside:
- Honest optimism. The planning fallacy at work. The team believed the number.
- Strategic misrepresentation. Flyvbjerg’s term for deliberately shading the estimate because you know every other proposal is shaded too. Nobody thinks of it as lying. They think of it as competing.
In most organizations both forces are present at once. That’s why the fix has to be structural rather than moral. You can’t exhort people out of a selection effect.
Why do projects go over budget?
Most overruns weren’t caused during the project. They were already in the business case on day one, undiscovered.
The baseline was set at the moment of maximum optimism: before anyone had done the work that reveals the truth, in a document written to win. Everything that follows is discovery, not failure. The vendor delay, the integration nobody scoped, the adoption curve that turned out to be a curve: each one is a fact the estimate never contained.
That reframe changes where you intervene. Fixing execution can’t recover an estimate that was wrong at approval. You can run the project perfectly and still land exactly where the honest number said you would.
| The overrun didn’t happen in month nine. It happened in the business case. Month nine is when you found out. |
Discerning the Signal in the Noise: Why It Matters for Project Managers
How do you recognize a cursed business case?
- Single-point estimates. One number, no range. A range would have shown how wide the uncertainty really was.
- Benefits that assume full adoption on day one. Real adoption is a curve, and the curve is where most benefit shortfalls hide.
- No reference class. Nobody asked how comparable projects actually finished.
- The sponsor wrote the estimate. The person who most wants the project approved produced the number that got it approved.
- Contingency was trimmed to make the number work. Usually late, usually the night before the review.
- The risk register only lists risks that were already known. Nothing about what would have to be true for this to be a bad idea.
- No downside scenario. A business case with only an upside is a pitch, not an estimate.
How to Use the 5 Whys Technique at Work (Step-by-Step With Examples)
What do you do if you’ve inherited one?
- Re-baseline early. The sooner you do it, the less sunk cost is standing in the way of an honest number.
- Convert points to ranges and report the range. A range tells the sponsor the truth without accusing anyone.
- Separate the base case from the optimistic case, and name which one the approval was actually based on.
- Find the reference class. How did the last ten projects like this one finish? Part 4 of this series shows the method.
- Escalate with data, not opinion. “Comparable projects finished at this range” lands. “I think this is tight” doesn’t.
- Insist on stage gates with real decision rights. A gate that can’t say no is a status meeting with a nicer name.
Frequently asked questions
What’s the difference between optimism bias and the planning fallacy?
The planning fallacy is the specific tendency to underestimate how long a task will take and what it will cost, even when you know similar tasks ran over. Optimism bias is the broader tilt toward favorable outcomes across cost, duration, benefits, and risk.
Crafting Tailored Project Plans that Deliver Real Business Value (and How AI Can Help)
Is strategic misrepresentation just lying?
It’s shading a forecast to win approval, usually because the competing forecasts are shaded too. The people doing it rarely experience it as dishonesty, which is why controls fix it and exhortation doesn’t: independent estimates, reference classes, and gates with the authority to say no.
How do you build a business case that isn’t optimistic?
Use ranges instead of points, base the estimate on how comparable projects actually finished, model benefits as an adoption curve rather than a switch, and include a written downside scenario. Then have someone who doesn’t want the project approved review the numbers.
| WHERE THE NEW PMP® EXAM TESTS THIS
Business case validation, benefits realization, and project selection live in the Business Environment domain, which is 26% of the PMP® exam since July 9, 2026, up from 8%. The new case-study question sets walk you through a project whose business case has drifted and ask what you do about it. The tested skill isn’t spotting that the numbers were optimistic. It’s knowing what to do once you have. |
| Business cases, benefits, and value are a quarter of the exam now.
Business case validation, benefits realization, and project selection sit in the Business Environment domain, which tripled on the PMP® exam that launched July 9, 2026. Our PMP® Certification Online Training Bundle covers all of it: 35 contact hours, 2,400+ practice questions, and 9 realistic exams, self-paced. |
NEXT IN THE SERIES
Part 3 — Unrealistic Project Deadlines: The PM Who Said Yes Fastest
PMI, PMP, PMBOK and PMI-ACP are registered marks of the Project Management Institute, Inc.