Where estimates leak money

A 12-metre ridge is not one piece of capping. Flashing comes in 7 and 8 metre stock, so that ridge is two pieces — and every joint eats a lap, typically 100 mm. Estimate the measured length and you are short one lap per joint, every run, every line type. On one roof that is trivial. Across a year of quotes it is a steady, invisible discount you are giving away.

Plug it: price from a cut list split to the stock lengths your supplier actually sells, with laps added per joint — not from the measured total.

Most quotes add a flat waste factor — “10% should cover it”. But cutting waste is not an opinion, it is an outcome of the nesting: which sheet lengths, cut where, for which planes. On a real hip-roof job we studied, the ordered sheet area came to 10.9% over the roof area — not because someone chose 10.9%, but because 72 specific sheets in 9 specific lengths is what covered it. A guessed factor is sometimes generous, sometimes short, and you never learn which.

Plug it: derive waste from the actual cut list where you can; where you must use a factor, record it per material and true it up against what the job actually consumed.

The price of steel, concrete and framing timber moves. A rate you keyed in eighteen months ago is not a price — it is a memory. The dangerous part is that a stale rate does not look stale in a spreadsheet; it looks exactly like a current one, right up until the supplier invoice lands.

Plug it: date every rate. Flag anything older than six months at quote time, and never let a quote silently pick up a price that was not current on the day you priced it.

Material is bought at the wastage-inflated quantity — you order 110 m² of board to line 100 m² of wall. Labour is not. Nobody stands there installing the offcuts. If your sheet applies the same inflated quantity to both columns, every line quietly overprices labour by the waste factor.

Plug it: cost material on the ordered quantity, labour on the measured quantity. Two columns, not one.

Supplier exports say LM, SQM, EA, CUM — and one day a square-metre rate lands on a lineal-metre line. The arithmetic still works. The total still foots. It is just wrong, by whatever the ratio of perimeter to area happens to be on that job. Unit errors are the hardest leak to spot because nothing about the quote looks broken.

Plug it: make the unit part of the line, not a habit of the estimator — and refuse to let a rate in one unit price a quantity in another.

Drawings move: REV A becomes REV D, the garage grows half a metre, a window becomes a door. The quantities you took off REV A do not know that. On volume-builder work, where the same plan cycles through revisions quickly, pricing a superseded sheet is one of the most expensive routine mistakes there is — because the contract references the drawing, not your takeoff.

Plug it: record the revision against every quantity, and when a new revision lands, list what it touches before anyone re-sends the price.

Ask estimators where quotes really go wrong and the honest answer is rarely the takeoff — it is the item with no line at all. The pump hire. The scaffold. The waterproofing angle fillet. A measured quantity can be 5% off and you survive; a missing scope item is 100% off, every time. This is the largest leak on this list and the least glamorous.

Plug it: price against a per-trade checklist of what the scope usually contains, so an absent item is a decision you made, not an accident you find in month two.

Measure a house by its walls and you will miss the roof. Porticos stand on posts. Alfrescos are open on two sides. Garage frontage is a door, not a wall. All of it is roofed, none of it is walled — and on a typical volume-builder home those areas add up to 20–30 m² of roof that a wall-derived area simply does not contain. We measured the gap on a real plan set: 11% of the roof, missing, silently.

Plug it: take the roof off the roof outline — eave line, not wall line — and treat any “area × factor” shortcut from the floor plan as an estimate of an estimate.

Six months after the job, a variation dispute lands and someone asks: which rate did this line use, from which price list, against which drawing revision? If the answer is “whatever was in the spreadsheet at the time”, you are negotiating from memory against their paperwork. An estimate that cannot be reproduced cannot be defended.

Plug it: keep the evidence with the number — the rate’s date and source, the drawing revision, the measured-versus-assumed status of every line. Boring on the day you quote; decisive on the day you argue.