Construction Cost Control: Why the Gap Between Quote and Spend Grows
Estimating gets the attention, but the money is usually lost in the space between a variation being agreed on site and being recorded in the office. That space is a data problem, and it is fixed by design rather than by discipline.
A contractor wins a project on a priced schedule, then spends the next eight months discovering that the schedule and reality disagree. By the final account the margin has moved, and the explanation is usually a list of variations that were verbally agreed, materials that were bought at a different rate than assumed, and labour hours that were never measured against anything.
The estimate is rarely the problem
Estimating errors are visible early and get corrected. The structural problem is that a construction project generates cost information at a rate the office cannot absorb: daily site activity, delivered materials, subcontractor progress, change instructions. If that information reaches the finance system a month late and in summary form, the project is being managed on history.
Three gaps where margin disappears
Gap one: the variation that was agreed but not written
A site instruction given on a Thursday and started on Friday is commercially agreed in practice and undocumented in the system. By the time it is formalised, the work is done, the quantity is a memory, and the negotiation is about recollection rather than record. The fix is procedural and technical at the same time: the site must be able to raise a variation in under a minute from a phone, with a photo, and that record must be visible to commercial staff the same day.
Gap two: materials bought against a rate that changed
Purchase orders are raised against the priced schedule, but the actual invoice arrives at a different rate or a different quantity. If the reconciliation happens monthly in a spreadsheet, the project carries an unknown exposure for weeks. The system requirement is simple: every purchase order line carries the budget code it consumes, and every invoice is matched to the order line rather than to the project as a whole.
Gap three: progress measured in percent rather than in quantity
"Seventy percent complete" is a judgement, not a measurement, and it is usually optimistic in the last twenty percent. Progress measured in installed quantity against priced quantity produces a number that can be checked and argued about with evidence. It also makes subcontractor valuations defensible.
What the system has to hold
| Record | Minimum fields |
|---|---|
| Priced schedule | Line item, quantity, unit rate, budget code |
| Purchase order | Budget code per line, supplier, expected delivery |
| Supplier invoice | Matched to order line, rate variance flagged |
| Variation | Location, description, photo, who instructed, date raised, priced status |
| Site progress | Installed quantity per line item, date, recorded by |
| Labour | Hours by task and crew, not just total hours on site |
The pattern is the same in each row: cost must attach to a line, and the line must attach to the estimate. A system that records totals per project cannot answer the only question that matters mid-project, which is where the money is going.
Designing for the site, not the office
Construction software fails when it assumes a desk. Site staff work with gloves on, in sunlight, with intermittent signal, and they will not fill in a twelve-field form to report a delivery. Design consequences:
- Capture offline and sync later; never block a record on connectivity.
- Photograph first, describe second — a photo with a code beats a paragraph.
- Default every field that can be defaulted from context: project, date, user, location.
- Keep offline mode honest — the record must be visibly pending until it syncs, not silently lost.
- Report exceptions rather than totals: what is over budget, what is unrecorded, what is awaiting pricing.
A staged scope that holds
- Stage one: priced schedule and purchase orders with budget codes — the structure everything else depends on.
- Stage two: site capture for deliveries and variations, with photo and offline sync.
- Stage three: invoice matching and rate variance reporting.
- Stage four: progress in installed quantity, subcontractor valuations and the final account.
The first stage is unglamorous and it is the one that determines whether the rest works. Without a priced schedule that matches how the project is actually bought and built, every later report produces numbers nobody can trace back to a decision.
Questions we are asked
Can this work with our existing accounting package?
Yes, and it should. Accounting keeps the ledger; the project system keeps the commitment and the evidence. The integration is limited to cost codes, supplier invoices and the periodic cost-to-date, rather than duplicating the ledger.
How do we handle subcontractors who do not use software?
They do not have to. The site team records the progress and the delivery; the subcontractor receives a valuation document produced from that record. The evidence lives on one side, which is also the side that needs it.
What if our projects are small and infrequent?
Then the cost of a bespoke system outweighs the benefit, and a structured spreadsheet with the same fields is a reasonable interim step. The design principles — cost attaches to a line, variations captured on site — still apply.
