A construction contract is signed on a price for a defined scope, and then the scope changes. That is not a failure of the process, it is how building works: ground is never exactly as drawn, clients change their minds, and designers issue information late.
The consequence is that a large share of the final account on a typical project is not in the tender sum at all. It is in variations, and the way they are instructed, valued and recorded decides whether the project is profitable.
This guide covers what counts as a variation, how to value one, how the monthly valuation works alongside it, and the record keeping that makes both defensible.
A variation is a change to the scope, the quantity, the quality or the timing of the work, instructed under the contract. In practice they fall into four groups:
The last two are where contractors leave money behind, because they do not look like variations on a drawing. They look like normal site difficulty, and they get absorbed.
Nothing is a variation until it is instructed. That rule protects both sides, and it is broken in the same way on almost every job: a verbal instruction from a client or a site manager, acted on immediately, documented weeks later or not at all.
The disciplined version is short. Confirm every verbal instruction in writing within a day, on the day, with a simple statement of what you understood was asked, what it costs and what effect it has on the programme. If the client disagrees, you find out before the work is built, which is the cheapest possible moment.
A written confirmation that goes unchallenged is the strongest evidence there is. A photograph of the finished work with a verbal story attached is the weakest.
Contracts normally set a hierarchy for valuing changes:
Dayworks is the fourth option, used where the work cannot be measured in advance, and it is priced from an agreed schedule of labour and plant rates.
The trap is the contract rate itself. A BoQ rate for a large quantity of plasterboard reflects the productivity of doing nothing else. Applying the same rate to 12 m² in a finished room, with furniture to protect and a working day that starts after the client has left, is not a similar character of work. That distinction is worth making politely and early, and it should be made before the work is done rather than at the final account.
A valuation is a measurement of work executed, priced at contract rates, plus a proportion of any variations and any amounts due under the contract for materials on site.
Two disciplines make a valuation survive scrutiny:
Measure the same way each month. Use the same method, the same units and the same level of breakdown as the priced document. A valuation that changes its basis is a valuation that invites a re-measure.
Do not claim what is not built. An optimistic valuation is a short term cash win and a long term credibility loss. A certifier who has caught an overclaim once will check every subsequent one closely, and the delay that follows costs more than the amount claimed.
Work executed but not yet measured, materials delivered but not yet fixed, and variations instructed but not yet valued all belong in a valuation, clearly identified. They are legitimate claims, and leaving them out is how a contractor finances the client's project for free.
Almost every construction contract holds back a percentage of each valuation, commonly 5 percent, half released at practical completion and the balance at the end of the defects period.
The retention is not a cost, it is a delay, and it belongs in the cash flow forecast from the first month. A contractor who forgets retention discovers in month four that the valuation is 100 but the payment is 95, every month, and the difference is the profit.
Set off is the other side of the same coin: the client's right to deduct for defects or for costs they have incurred. It should be notified with a reason and an amount. An unexplained deduction is a dispute, and it should be answered in writing within the time the contract allows.
Six months after the event, the only facts that exist are the ones that were written down. The record set that matters is small and cheap:
The site diary is the single most valuable document on a project and the one most often skipped. It costs five minutes a day and it is the difference between a claim that is paid and a claim that is negotiated down to nothing.
Time is money on a construction site, because the preliminaries run per week whether or not work is progressing. A four week delay caused by late information is a quantifiable loss, and it is recoverable if the contract allows and the records exist.
The evidence required is a chain, not a story:
Claims fail on the third link far more often than the first. Everyone remembers the delay. Fewer can demonstrate that it moved the completion date rather than being absorbed by float elsewhere.
Most variation disputes are not about money. They are about surprise.
A client who is told on day three that moving the door will cost 900 and add two days can decide whether they want it. A client who discovers the same figure on the final invoice feels deceived, even though the work was done properly and the price was fair.
That is the practical argument for valuing before building, and it works in the contractor's favour in a way that is easy to miss. A client who has approved five variations through the job expects a final account that reflects them. The final account then arrives without drama, the retention is released without argument, and the recommendation to the next client is made willingly.
On a project with several trades, one instruction usually affects more than one of them. Moving a wall changes the plaster, the electrics, the flooring and the decoration, and each of those is somebody's variation.
The coordination point is that the instruction should be passed down to the affected trades on the same day, with the same scope description. If each trade describes the change differently, the client receives three documents that do not reconcile, and the whole variation process loses credibility.
A single variation notice, listing each trade's element and each trade's price, is more work on the day and far less work at the final account.
A kitchen refurbishment, contract sum 24,000, eight weeks.
Instructed variations during the work:
Total variations 3,810, which is 16 percent of the tender sum. On a job of this size that proportion is normal, and it is why the final account is so often materially different from the tender.
The point of the example is not the money. It is that the worktop change was instructed in writing on day nine, the rotten floor was photographed and signed off the same afternoon, and the delay was recorded in the diary with the date the information was chased. All four survived to the final account without argument.
Not every variation adds money, and the ones that reduce it are the ones contractors forget to value.
An omission should be valued on exactly the same basis as an addition: at the contract rate, less the costs that are genuinely saved. Those costs are rarely the full rate, because the preliminaries, the supervision and the overhead do not reduce with the work. A wall removed from the scope saves the plastering but not the scaffold that was erected for the whole elevation.
Valuing omissions fairly, and saying so, buys an enormous amount of credibility for the additions that follow. A client who sees a reduction calculated honestly is far more likely to accept an increase calculated on the same basis.
The final account is the reconciliation of everything that happened, and it should be built from documents rather than from memory.
The version that closes quickly has four parts: the measured work at contract rates, the variations listed with their instruction dates and references, the adjustments for provisional sums, and the deductions for retention and any set off. Each part is a page a client can check against their own file.
The version that drags on for a year is an aggregate figure with a narrative attached, and it is usually the contractor who suffers, because the burden of proof sits with the party asking for the money.
What is the difference between a variation and a claim? A variation changes the scope and is instructed. A claim is for money or time arising from an event, such as a delay or a breach, rather than from a change in the work.
Can I refuse to do a variation? Usually not if it is validly instructed under the contract, but you can price it, and you can state its effect on the programme.
Should variations be valued before the work is done? Wherever possible. It removes the argument at the final account and it gives the client a decision to make.
What is dayworks? Work priced from an agreed schedule of labour, plant and material rates rather than from measured quantities, used when the work cannot be measured in advance.
How long do I have to notify a delay? It is set by the contract, often 14 or 28 days from awareness. Missing it can extinguish the entitlement entirely.
Why is the final account always higher than the tender? Because scope changes, and because the tender sum is a price for the original scope rather than a forecast of the final cost.
The FRELUX project tools record each variation with its instruction date, its valuation basis and its programme effect, and roll them into the monthly valuation so the final account is built as the work proceeds rather than reconstructed afterwards.