Setting a Renovation Budget That Holds

The three buckets, and what falls between them

Materials, labour, and permits and fees. It looks like a trivial split until you try to file real quotes into it and find that a third of them straddle two buckets. A fitted-kitchen supply-and-install price is both. A tiler charging by the square metre and supplying adhesive and grout is mostly labour with a material tail. Decide on a rule before you start — the cleanest is that anything on a supplier invoice is material and anything on a trade invoice is labour, with supply-and-fit prices going wherever the larger half sits — and then apply the same rule to every line, because a budget that mixes conventions cannot be compared with anything, including its own earlier version.

Labour carries a question worth asking out loud: is the figure a fixed price for the work, or a day rate against an estimate of days? A fixed price already has the contractor's overhead and profit inside it, along with whatever they privately allowed for the job running long. A day rate does not, and it transfers the risk of a slow week from them to you. Neither is wrong, but they must not be added to one total without noting which is which. The National Association of Home Builders publishes its Cost of Doing Business Study on exactly this ground, and it is a reminder that a builder's margin is not a discount waiting to be found.

Permits and fees is the bucket that is almost always too small, because most people put a permit fee in it and stop. Building permits are only the first line. Section R105 of the International Residential Code is what requires the permit for residential work in adopting jurisdictions, and R105.2 is the short list of work exempt from one — worth reading, because it is shorter than people assume. Around that sit plan review fees, per-visit inspection charges, planning or zoning consents where the outside of the building changes, utility disconnection and reconnection charges, a highway licence if the skip or the scaffold stands on a public road, and in some jurisdictions a party wall award before you touch a shared structure. None of these are large individually. Together they are routinely a four-figure sum nobody had a line for.

Then there is the material that is not part of the finished house and therefore lands nowhere: floor protection and dust sheeting, tool and plant hire, a temporary kitchen, storage for what comes out of the rooms, container hire and tipping charges, and making good the parts of the house the works pass through rather than change. On a domestic job with people still living in it, that last category is not small. Put it in the materials bucket by name rather than letting it be absorbed as a surprise later.

1. Fix the bucket rule first — supplier invoice is material, trade invoice is labour — and write it down where you will see it when a supply-and-fit quote arrives.
2. Enter material figures at exact measured quantity, with no overage in them yet — buying it twice is the most common arithmetic error in a domestic budget.
3. Enter labour as quoted, and flag every line that is a day rate rather than a fixed price. Those lines are the ones your contingency is really covering.
4. Ring the building department and the utilities before you fill the permits bucket. Fee schedules are published, local, and nothing like each other.
5. Add the non-finished material — protection, hire, storage, disposal, making good — as named lines rather than a round sum, so that later you can tell which of them you actually spent.
6. Only now choose the contingency percentage, and choose it against the register of unknowns rather than by habit.

Overage is bought, not wasted

The word waste does the damage. It suggests carelessness, when in fact the offcut in the skip is a quantity you deliberately purchased so the room could be finished. A rectangular room tiled in a plain square format wastes very little. The same room in a herringbone lay, or with a diagonal set, or with a pattern that has to match across a seam, throws away a materially larger share before anyone has made a mistake. That share is a property of your geometry and your chosen material, not of the fitter, and it can be reasoned about before anything is ordered.

There are two places to add it, and doing both is the classic double-count. Some quantity calculators on this site carry a waste field of their own — you give it a floor area and a percentage and it returns boxes with the allowance already inside — while most return the exact quantity the geometry demands and nothing more. Which kind you used decides everything that follows, so look before you add anything: apply a cost-side percentage on top of a quantity that was already inflated and you have bought the overage twice. The cost-side buffer belongs where you hold a price for a precisely measured exact quantity, typically a supplier quote taken off your own measured schedule, and you need to know what the buying decision costs before you make it.

What moves the figure is worth being specific about, because the drivers are visible on the drawing rather than discovered on site. Module size against room size decides whether the last row is a full board or a sliver. Cut pattern decides whether the offcut from one row starts the next. Pattern repeat and dye lot decide whether an offcut is usable at all. Number of internal corners, alcoves, and pipe penetrations decides how many part-pieces exist. And the fitter's experience decides how many pieces are cut twice — a real factor on a first DIY attempt and an unfair assumption to make about a specialist.

No standard publishes an overage percentage, and any figure quoted as though one does should be treated with suspicion. The documents that govern these materials govern the installation, not the buying: the TCNA Handbook for Ceramic, Glass, and Stone Tile Installation sets out methods and substrate requirements, Gypsum Association GA-216 covers the application and finishing of gypsum panels, the NRCA Roofing Manual covers roof assemblies. Where a manufacturer directs an overage it is in their own installation instructions for their own product, and that instruction is the one to follow, because it is the one tied to the warranty. Beyond that, the only trustworthy number is the one you produce yourself by reconciling what was delivered against what was fixed on a job you have already done.

One argument against trimming the overage to nothing: leftover material is not entirely a loss. A box of the same tile from the same batch, a spare length of the same board, a litre of the mixed paint colour, all have real value the first time something is damaged and an exact match is needed. Attic stock is cheap when it is bought with the original order and effectively unobtainable two years later.

Sizing the contingency against named unknowns

Ten to twenty per cent is the range everybody quotes, and as a decision it is worthless, because the difference between the two ends of it is the difference between a project that survives one structural surprise and one that does not. AACE International Recommended Practice 40R-08, Contingency Estimating — General Principles, is clear about what the pot is for: uncertainty within the defined scope, the things experience says will turn up, priced as an amount likely to be spent. It is not for scope you decide to add later, and it is not a discount you get to keep if you are careful. Recommended Practices 41R-08 and 42R-08 set out the range-estimating and parametric methods for larger work; on a house the same logic fits on one page.

Three pots get confused with each other and should not be. Contingency covers unknowns inside the agreed scope. An allowance is a placeholder for a scope item that is definitely happening but has not been chosen yet — the tile you have not picked, the lighting you have not specified. A reserve, in the sense the PMBOK Guide separates management reserve from contingency reserve, is money held outside the project for scope changes. RICS New Rules of Measurement 1 does the most useful version of this split for building work, dividing the risk allowance into four named pots — design development, construction, employer change, and employer other — which forces the question of whose risk each pound is covering. That question is the whole exercise.

Build the percentage from the bottom rather than picking it from the top. Write out every unknown you can name, put a rough cost against the version of it you would actually have to fix, and weigh each by how likely it is. An unopened floor in a house of a certain age, a consumer unit with no spare ways, a flat roof of unknown date, a soil stack in the wrong place, a single-quote trade with no second price, a long-lead item that may be discontinued: each of those is a line with a number. Sum the weighted exposures, express that as a share of the base budget, and you have a percentage with a paper trail behind it. Nine times out of ten it lands inside the range everybody quotes, but you now know why, and you know which unknown to spend a survey on to bring it down.

What pushes it up is specific and mostly knowable in advance: the age of the building, whether anything at all has been opened up, whether the existing drawings agree with the tape measure, how many trades are priced from a single quote, how much of the programme runs while the house is occupied, and how long the gap is between the quotes and the start on site. That last one is escalation rather than risk, and it is worth carrying as its own line rather than hiding it inside the contingency percentage, because it behaves differently — it grows with time whether or not anything goes wrong.

What brings it down is only ever investigation. There is no other lever. Opening up a section of floor, having a structural engineer look at the wall you intend to remove, getting a second price on the largest trade — each of those retires a named risk and lets you defend a lower number. Cutting the percentage because the total came out uncomfortable is not a reduction in risk, it is a decision to fund the risk from somewhere you have not identified yet.

1. List the unknowns by name. If a line reads "something might go wrong upstairs", it is not finished — say which thing, in which room.
2. Price each one at what fixing it would actually cost, not at what you hope it costs.
3. Weight each by likelihood, honestly. A near-certainty is not a risk, it is a scope item that belongs in the base budget.
4. Sum the weighted exposures and express the total as a percentage of the base. That is your starting figure.
5. Test it against the top three lines. If retiring one of them by survey costs less than the exposure it carries, buy the survey and rerun the register.
6. Write the register down. It is what tells you, four weeks in, whether a bill is a draw on contingency or a scope change you chose.