April 2, 2026

How much contingency is enough?

Ask three people what contingency to carry on a construction project and you’ll get three defaults: 5% because that’s what the pro forma had room for, 10% because that’s what everyone says, or a single blended number nobody can explain. All three miss the point.

Contingency isn’t a convention. It’s a measurement of what you don’t know yet — and what you don’t know changes over the life of a project. Sized properly, contingency is the most honest line in a budget. Sized by habit, it’s the first place a project starts lying to itself.

Why one blended number fails

A single “10% contingency” line hides four different risks that behave in four different ways. When they’re blended, you can’t tell whether the money is being consumed by design growth (normal, should be shrinking), site surprises (concerning, should be investigated), market movement (external, should be tracked), or scope creep (self-inflicted, should be governed). The total just goes down, and nobody learns anything until it’s gone.

We carry contingency in named buckets on every project we manage, because each bucket tells you something different when it moves.

The four buckets

Design contingency covers the drawings maturing — the difference between what’s drawn today and what a complete, coordinated set will require. It should be largest at concept (when the drawings are mostly promises) and approach zero at tender (when they’re documents). If design contingency isn’t shrinking as design progresses, the design isn’t actually progressing — it’s growing.

Construction contingency covers what the ground and the building hide: soils, water, existing conditions, coordination collisions. Unlike design contingency, it doesn’t burn off before construction — it exists for construction. Renovation and heritage work needs materially more, because old buildings keep secrets until you open them up.

Escalation covers the market moving between budget date and purchase date. It’s technically not contingency — it’s a forecast — but it belongs in the risk conversation because it’s uncertain, directional and large. It should be carried to the schedule you actually believe, not the one in the brochure.

Owner contingency covers you changing your mind — which you will, and which is your right. Separating it matters because it’s the only bucket entirely within your control, and the only one it’s legitimate to spend on making the building better rather than making problems go away.

Benchmarks by stage

Every project is different, but as honest starting points for total carried risk (excluding escalation):

• Concept / feasibility: 12–20%. Mostly design contingency. If your concept budget carries 5%, you don’t have a tight project — you have an undeclared risk.
• Schematic design: 10–15%, design bucket starting to shrink as decisions land.
• Design development: 8–12%, with construction contingency now explicit.
• Tender, on complete documents: 3–7% construction contingency — the low end for new build on a clean site with thorough investigation, the high end for renovation, heritage, or fast-tracked documents.

Fast-track delivery deserves its own warning: if you tender from incomplete documents, the design contingency you “saved” hasn’t disappeared. It’s moved into the contractor’s change orders, at site prices. We cover that mechanism in where change orders actually come from.

Managing the drawdown

Contingency only works if the drawdown is controlled. The discipline is simple and rarely followed:

Every use gets logged — what it paid for, which bucket it came from, why the item wasn’t in base scope, and what remains. Drawdown is approved, not absorbed — someone with authority says yes before the bucket shrinks, rather than discovering it shrank. The burn rate gets reviewed monthly against progress — 60% of construction contingency spent at 30% construction complete is a projection, not a coincidence, and it’s telling you the end of the story early. And nobody parks scope in it — the moment contingency becomes the place unfunded wishes go, it stops being risk money.

When contingency is a slush fund, it disappears without a lesson. When it’s a ledger, it becomes the most educational document on the project — a running record of what the team didn’t know, and when they found out.

What healthy looks like

On a well-run project, the contingency story arcs predictably: broad at concept, burning down through design as decisions replace assumptions, stable through tender, then construction contingency depleting roughly in step with construction progress — faster in the ground, slower in finishes. A project tracking that arc is learning at the rate it’s spending. A project off that arc has something to say, and the sooner it’s heard the cheaper it is.

Setting those buckets, benchmarking them against the project’s real risk profile, and policing the drawdown is core preconstruction planning work — and one of the first things we look at when we’re brought in to review a budget someone else built.

Frequently asked questions

Is contingency the same as an allowance?

No. An allowance is a placeholder for a known item not yet designed — a lobby finish, a signage package. Contingency covers unknowns. Projects get into trouble when allowances are quietly under-scoped and contingency absorbs the difference, muddying both.

Should the contractor hold contingency too?

They do, whether it’s visible or not — inside their price, inside their schedule. On CM and open-book arrangements, insist on seeing it: whose risk it covers, who approves its use, and where unspent amounts go.

What happens to unspent contingency?

Decide before the project starts, in writing. On owner-held contingency the answer is simple — it’s your money. On shared or contractor-held pots, silent assumptions about “savings” cause end-of-job fights that a paragraph in the contract would have prevented.

If your current budget carries one blended number and you’re not sure what’s inside it, that’s a one-week review — and it’s far cheaper before tender than after.