How much contingency a renovation needs, and the rules for spending it
Hold 10 to 15 per cent of the build cost as contingency on a substantial UK renovation, which is the range most 2026 guides recommend, and push towards the top of it or beyond when the building is old, the ground is unknown, or the work opens up existing structure. That is the short answer. The useful part is the rules around it, because contingency fails less often by being too small than by being spent on the wrong things.
What contingency is for
Contingency exists for discovery: the costs that could not have been known when the budget was written. The floor that comes up rotten. The wiring that turns out to predate colour television. The ground that needs deeper foundations than the trial hole suggested. These are not planning failures, they are the irreducible unknowns of working on an existing building, and the older and less documented the building, the larger they run.
What contingency is not for
It is not for upgrades. The moment contingency starts buying nicer taps, better tiles or the underfloor heating that did not make the original budget, the project is spending its insurance on treats, and the rot under the floor is now unfunded. Upgrades are a legitimate choice, but they are a scope change, and a scope change should be a conscious decision against the budget, not a quiet draw against the reserve.
It is also not padding hidden inside individual lines. Contingency inflating every line by a little disappears into the quotes; held separately and visibly, it is a fund with rules.
Sizing it
Ten per cent suits a well-scoped project on a structurally understood building with drawings and a full specification. Fifteen per cent suits Victorian and older housing stock, projects that open up floors, roofs or chimney breasts, and anywhere the survey stopped at the wallpaper. Beyond fifteen belongs to genuine unknowns: basements, underpinning, buildings with a history of unrecorded alteration. A new-build garden room on a clear site can justifiably run below ten.
Two things move the number more than any rule of thumb: how much investigation was done before pricing, and how much of the existing building the work touches. Money spent on trial holes and a proper survey is usually cheaper than the contingency it lets you not hold.
The rules for releasing it
Release contingency when something is discovered, not when something is desired. A release should name the discovery, the cost, and what remains in the fund, in writing, even if only to yourself, because a fund drawn without a record empties without a reason. If the fund passes half spent before the build passes half done, that is the earliest reliable warning the project has moved from unlucky to underpriced, and the response is a re-forecast rather than hope.
And if it is unspent at the end, it was never a savings target. It goes back where it came from.
Common questions
Is contingency on top of the quotes, or inside them?
On top, always, and visibly. Builders' quotes carry their own risk pricing inside them; your contingency covers what no quote has priced because nobody has seen it yet.
Should contingency cover price rises during the build?
On a project of a year or less, mostly no: inflation belongs in the estimates themselves, and UK construction output prices have moved sharply in both directions within single recent years, so a static allowance dates fast. On multi-year projects, hold a separate, named inflation allowance rather than letting it eat the discovery fund.
The builder says his price includes contingency. Does that count?
No. His contingency protects his margin on his scope. It does not cover the rotten joist he has not seen, and it evaporates entirely if the discovery becomes a variation order, which it will.