In this article
You have a quote, a schedule, and an opening date you've already told your staff. At least one of the three is almost certainly wrong — not because your contractor lied or you were careless. There's a name for what's happening, and a way to correct for it in advance.
Every owner who has ever renovated knows the feeling: the quote said €45,000, the invoice came to €60,000. The schedule said ten weeks, the doors opened after fifteen. It's easy to blame the contractor, the permit office, or the supplier who delivered late — often fairly — but there's a pattern underneath all of it that has had a name since 1979.
Daniel Kahneman and Amos Tversky called it the planning fallacy: people systematically underestimate how long and how much a project will take, even when they know full well that similar projects have run long before. It doesn't happen because people plan badly — it happens because they plan from the inside view (this specific project's own steps) instead of the outside view (how similar projects actually turn out in practice).
The inside view feels reliable precisely because it's so concrete: you can see the steps, you know the contractor, you're holding the quote. The outside view feels vague — "projects tend to run over" — and gets systematically ignored, even by people who know the statistic.
This guide walks through seven numbers that together show how predictable this bias really is — from where the concept comes from, through a dataset of 258 infrastructure projects and a figure specific to restaurant renovations, to the maths that explains why a project with several sequential steps runs later than intuition suggests. At the bottom, run your own renovation through the same method a real government department already applies.
Why even experienced planners keep making this mistake
The inside view almost always beats the outside view, because it's the only information that's readily available at the moment you're planning. You're looking at THIS quote, THIS contractor, THIS permit application — not at a dataset of hundreds of other renovations you don't actually have in front of you.
There's a social layer on top of that: a contractor who honestly says "budget fifteen weeks and a 30% margin" loses the job to whoever promises "ten weeks, exactly on budget". Optimistic schedules aren't just made — they're rewarded, right up until the bill arrives.
And the bias doesn't fix itself with experience. Someone renovating their second restaurant remembers exactly how the first one ran over — and still underestimates the second time, because every new project feels like a unique case rather than the next instance of the same category.
The ultimate guide The ultimate guide to restaurant finance From food cost to cash flow: every number that steers your business, in one guide. Open the guideThe 7 numbers, and what each one tells you
They're ordered the way the argument builds: first where the bias comes from, then how big and how universal it is, then the maths that explains why it's so persistent, and finally the method that already exists to correct for it.
1. 1979 — the year the mistake got a name
Kahneman and Tversky first described the planning fallacy in their paper "Intuitive prediction: biases and corrective procedures" (TIMS Studies in Management Science, 1979, vol. 12, pp. 313–327). Their core point: people predict how long something will take by imagining that specific project's own steps — the inside view — and systematically ignore how comparable projects actually turned out — the outside view.
The sting is in the word "systematically": this isn't a knowledge gap. Ask someone directly about their last renovation and most will admit it ran over too. That same person will still predict a schedule with no margin for the next one — because every new project instinctively feels like the exception, never the repeat.
That's exactly why "just plan better" doesn't work as a fix: the problem isn't in the details of the plan, it's in the perspective the plan is made from. The correction later in this guide (the outside view, applied numerically) is therefore not a better plan — it's a different kind of plan.
2. 9 in 10, +28% on average — the dataset that makes the pattern visible
Bent Flyvbjerg analysed 258 major infrastructure projects (rail, roads, bridges, tunnels) across 20 nations on 5 continents, worth roughly $90 billion, spanning seventy years. Result: 9 in 10 projects ran over budget, by an average of 28% — rising to 45% for rail projects specifically.
The most unsettling detail isn't the percentage — it's the timeline: that 28% overrun stayed roughly constant across seven decades. Better software, better project-management methods, more experience — none of it moved the pattern in decades, which strongly suggests the cause isn't in execution, it's in how the budget was estimated in the first place.
That's the core argument of this guide: if better tools and more experience don't fix it, the fix isn't a better plan from the same perspective — it's a different perspective. See number 5 for the method that came out of exactly that observation.
3. 34% — the number specific to restaurant renovations
Back to hospitality: independent industry surveys of restaurant renovation projects show an average budget overrun of 34% — higher than Flyvbjerg's general infrastructure average of 28%. The most commonly BUDGETED buffer in those same surveys is 10–20%, structurally too thin for what actually happens.
Why does a restaurant renovation fare worse than the average construction project? Two reasons that reinforce each other: an existing building only reveals its problems once the walls are open (wiring, plumbing, damp damage nobody could see before demolition), and a hospitality venue has a food-service permit, a fire-safety inspection, and sometimes a liquor licence on top of the building permit — each its own track with its own waiting time.
See our guide on restaurant renovation costs for what a make-over itself costs, and on restaurant permits for the tracks that usually cause the delay, rather than the building work itself.
4. 0.85⁶ ≈ 38% — the maths nobody runs, and why independent steps compound faster than intuition says
Here's the core reason the planning fallacy is so persistent, and it's pure arithmetic: if a renovation has six sequential, dependent steps — permit approved, contractor starts, structural work done, equipment delivered, inspection passed, opening — and each step individually has a comfortable 85% chance of landing on time, the chance that ALL SIX land on time isn't 85%. It's 0.85 raised to the sixth power: roughly 38%.
Every step looks completely safe on its own — 85% isn't a number that keeps anyone up at night. But because a renovation is a CHAIN (the inspection can only happen after the structural work, the opening only after the inspection), they all have to succeed at once, and probabilities you have to win one after another get MULTIPLIED, not added. That's exactly why a project with more moving parts runs later than the intuition of someone looking at each step in isolation would suggest.
The graphic below makes that tangible: six realistic, individually credible odds that together land well under a coin flip — not because one link fails, but because they all have to succeed simultaneously.
Three reference points on the same scale: your own plan as the baseline (100), Flyvbjerg's infrastructure average, and the hospitality-specific figure. Both sit well above what most owners plan for as a buffer.
These are averages across hundreds and 258 projects respectively, not a prediction for your specific renovation. The point of this graphic isn't the exact figure — it's the gap between what you're budgeting and what the reference class has shown for decades.
5. 4% to 51% — the methodology a real government department already uses to correct for this
The UK's HM Treasury publishes a supplementary guidance document to its own "Green Book" on what it calls "optimism bias": a table of mandatory capital-cost uplifts by project type, designed to correct for exactly this bias before a public project is approved. For non-standard buildings, that uplift ranges from 4% (lower bound) to 51% (upper bound) — and the full upper-bound uplift is applied mandatorily at the earliest business-case stage, only reducing once risks have been demonstrably identified and mitigated, never before.
This is called reference class forecasting: instead of estimating your own project from the inside view, you find a group of comparable projects (the reference class) and use their average overrun as a correction to your own estimate. It's the exact method Flyvbjerg himself helped develop, now turned into a mandatory government process.
For a restaurant renovation, the reference class isn't abstract — it's the figure from number 3 above (34%) and the general infrastructure average from number 2 (28%, up to 45% for the heaviest projects). The calculator further down applies exactly the same logic — a bracket per project type, applied to your own budget.
Each step carries an individual 85–95% chance of landing on time. Because the steps have to succeed one after another, the odds multiply — and the cumulative probability drops faster than intuition suggests.
These percentages are illustrative — every individual step looks credibly safe. The point is the multiplication: with six steps like these, the chance that ALL of them land on time drops well below a coin flip, without any single link needing to be "late" on its own.
6. 10–20% vs. 20–51% — the gap between what you budget and what the reference class says
Put the previous two numbers side by side and the gap is immediately visible: most owners budget a 10–20% buffer "for the unexpected". The reference class — Flyvbjerg's 28% average, the hospitality-specific 34%, and the Green Book's own 51% upper bound for non-standard buildings — says structurally more.
That isn't an argument for simply using the highest bracket. A light refresh (paint, furniture, lighting) deserves a smaller margin than a structural renovation with demolition and new plumbing — exactly why the calculator below uses a different bracket per project type, the way the Green Book itself does.
The point is that the buffer should be a deliberate choice based on a reference class, not a round number that "feels safe". 15% feels just as safe as 30% to most people, and yet the difference between the two is often the difference between a renovation you can finish and one that stalls halfway through for lack of money.
7. What one extra week actually costs — where the psychology turns into money
A schedule overrun isn't just an inconvenience — it's closed doors that could have generated revenue. Every week the opening slips is a week of revenue you don't make, while rent, the loan, and usually a chunk of your staff costs keep running regardless.
That figure becomes tangible once you combine the probability from number 4 with your own weekly revenue: at a compounding probability of 38%, you should statistically expect not your planned ten weeks, but ten divided by 0.38 — just over 26 weeks. That difference, sixteen weeks, multiplied by what an average week normally earns you, is what the planning fallacy quietly costs you — on top of the budget overrun itself.
That's exactly what the calculator below computes with your own numbers. Owners planning an opening usually look only at the cost of the renovation itself — never at the cost of the delay the renovation is likely to run into. Read more on what a closed week costs you in our guide on restaurant cash flow management, and on financing those first months in our guide on restaurant financing.
Run your own renovation through the outside-view method
Enter your own planned budget and schedule, pick the project type, and set how many sequential milestones your renovation has and how likely you rate each one individually. The calculator applies the same reference-class logic as the UK's Green Book, and the same multiplication logic as number 4 above.
The fields are pre-filled with a realistic example so you can see immediately how the result reads — adjust them to your own situation.
The outside-view calculator
Your budget and schedule, corrected with the reference class.
Project type
—
This is an illustrative model, not a quote or a guarantee: it assumes independent probabilities per milestone and a fixed bracket per project type. Your own permit process, contractor and building determine the real numbers. Everything runs in your browser; nothing is sent or stored.
The number that matters most here isn't the exact percentage — it's the gap between what you feel going in ("this will work out") and what the reference class has shown for decades. That gap IS the planning fallacy, made tangible in euros and weeks.
Use this before you sign, not after: once the contractor has started and the first invoices are in, the room to adjust your budget and schedule is far smaller than it was going in.
What to do with this before you sign the contractor
You don't correct the planning fallacy by trying harder — you correct it by deliberately planning from the outside view, at three concrete moments.
Before you sign a quote
- Decide which project type you have on the scale above, and use its bracket as your REAL budget — not as a worst case.
- Ask your contractor explicitly about the last two or three comparable projects they did, and how much longer and more expensive those ran versus the first quote.
- Count the number of sequential, dependent milestones in your own project, and run the multiplied probability through the calculator above yourself.
When building your permit schedule
- Ask your local authority for the ACTUAL processing time of your building, food-service, and (where applicable) liquor licence — not the statutory minimum.
- Schedule the food-service and fire-safety inspection in parallel with the final construction phase rather than after it, wherever that's legally possible.
- See our guide on restaurant permits for which tracks usually carry the longest wait.
When budgeting your opening
- Fold the calculator's expected delay cost into your financing plan, not just the renovation invoice itself.
- Build the reference-class buffer in as its own budget line, visible to you and your lender — not buried inside a rounded total.
- Repeat this exercise for every future renovation: the planning fallacy doesn't go away with experience, only with a deliberate correction.
The mistake isn't in your maths — it's in the perspective you're doing the maths from
Nobody who builds a renovation budget is being deliberately optimistic. The planning fallacy works precisely because it's invisible from the inside: every individual piece of the plan looks reasonable, and yet the whole thing systematically runs over — for you, for 9 in 10 major infrastructure projects, and for the average restaurant that renovates.
The correction isn't about planning harder — it's about using a different perspective: not "what's going to happen to me", but "what usually happens to projects like this" — and building that number in as a buffer up front, rather than experiencing it as a surprise later.
Put that reference class next to your own quote before you sign. See our guides on restaurant renovation costs and the startup budget & financing plan tool to map the full picture — investment, renovation, and buffer — before you begin.