Set Aside, Not Settled: A Commercial Discipline for Provisional Sums

Few lines in a UK construction contract carry as much quiet optimism as the provisional sum. It is a placeholder for work that is not yet fully defined, priced or designed. Yet once the tender is accepted and the number is transferred into the client's affordability model, it acquires a false solidity. The allowance becomes a budget line, the budget line becomes a cash-flow assumption, and the assumption becomes a commitment. The gap between what was set aside and what is eventually spent is then discovered late, when the commercial room to manoeuvre has narrowed.

That is the allowance illusion. It is not caused by dishonesty or incompetence. It is caused by treating a question as though it were an answer.

An estimate by another name

Provisional sums exist because not every element of a project can be defined at the point of contract. Under common UK forms, the distinction between a defined provisional sum and an undefined provisional sum matters. A defined sum suggests that the contractor has some information on which to base programming and preliminaries. An undefined sum carries less certainty still. In both cases, however, the sum is an allowance, not a fixed price.

The commercial difficulty is that tender documents, approval papers and funding models often present the total contract sum as a single figure. The provisional element is buried within it. By the time the project is underway, the allowance is rarely discussed as an estimate with a range of possible outcomes. It is discussed as money that exists. That shift in language is the beginning of the problem.

Why the number feels firmer than it is

Several pressures conspire to harden a provisional sum. Tender evaluation tends to reward a clean bottom line. Internal approval committees prefer a single number to a range. Project cash flows are built from monthly profiles that assume the allowance will be spent evenly or at a predictable point. And contractors, quite reasonably, plan their resources on the basis that the work will proceed.

None of these pressures is improper. Together, however, they create an expectation that the allowance will be spent, and spent within the sum. When the scope is finally defined, the conversation is no longer about whether the estimate was adequate. It is about who should have known, and who should pay for the difference.

Start with an assumptions schedule

The first practical step is to stop treating the provisional sum as a single figure and start treating it as a small portfolio of unresolved decisions. A provisional sum register should sit alongside the contract sum analysis and be reviewed at every commercial meeting.

For each sum, record the reference, the description, the amount, whether it is defined or undefined, the assumptions on which the allowance is based, the current design status, the decision owner, the target date for instruction, the amount drawn down to date, the remaining balance and a simple risk rating. The register should also carry a short narrative for each item explaining what would need to be true for the allowance to be sufficient.

This is not bureaucratic excess. It is the minimum evidence needed to answer a straightforward question: what did we think this would cost, and why?

Drawdown is a decision, not a deduction

A provisional sum should not be depleted simply because an invoice arrives. Every use of the allowance should be triggered by an instruction, a quotation or an agreed assessment, and a formal drawdown entry. The sequence matters: define the scope, obtain the price, instruct the work, then record the drawdown.

Where urgency requires work to start before the paperwork is complete, the confirmatory instruction trail should be closed the same day. Retrospective allocation is one of the most reliable ways to lose control of a provisional sum, because it converts a commercial decision into an accounting exercise.

Instruction before spend

The commercial value of a provisional sum depends on the discipline that surrounds it. No work order should be issued without a documented instruction that references the relevant sum. The authority to instruct should be clear, and it should sit with the person who understands the commercial consequence, not simply the person closest to the site.

This is not about slowing projects down. It is about ensuring that when money is committed, the commitment is visible. A provisional sum that is spent without an instruction is a variation waiting to be argued about.

Report the gap early

A monthly variance report should show, for each provisional sum, the original allowance, the amount instructed, the forecast to complete, the variance and a confidence rating. Where the forecast exceeds the allowance, or where the target instruction date has slipped, the item should be escalated.

Slippage is not only a programme issue. It is a commercial signal. A provisional sum that has not been instructed by its target date is likely to be instructed later, at a price that reflects a shorter period for pricing and delivery. The earlier that is visible, the more options remain.

Close the loop

At completion, every provisional sum should be reconciled formally. Unused allowance should be returned through the correct contractual mechanism, not left to be absorbed into the final account by inertia. The register should be closed with a short record of what was assumed, what was instructed and what was learned.

The provisional sum is not the answer to an undefined scope. It is the question. Treating it as a question, with a paper trail and a decision date, is how clients and contractors avoid discovering the answer too late.