Construction change control is usually presented as a procedure with forms and registers. It is better understood as four habits, and a project that keeps all four rarely ends in a claim.
Change itself is not the problem. Every project has it, and on most schemes it is a sign that decisions are being taken rather than avoided. The problem is change that is executed before it is valued.
One, identify it as change
Somebody has to notice that what is being asked for is different from what was priced. That sounds trivial and it is where the process most often fails, because the difference is frequently small and the person noticing is busy.
The practical test is whether the request corresponds to something in the contract documents. A drawing revision that adds an element, a specification substitution, a sequence alteration or a client preference expressed on site are all change even when nobody calls it that.
Scope creep is the accumulation of these. Individually each is too small to raise; collectively they are a substantial figure that appears at the final account with no owner.
Two, instruct it properly
A change becomes chargeable when it is instructed by the person the contract names, in the form the contract requires. Anything else is work done on the strength of a conversation.
Verbal instructions are the standard failure, and the standard fix is confirmation in writing on the day, every time, by whoever received it. That habit is covered in instructions given verbally and paid for twice.
Establishing at the pre start meeting who can instruct, and circulating that in writing, takes ten minutes and removes an entire category of argument.
Three, value it before it is executed
This is the step that separates a controlled project from an uncontrolled one. A change priced before the work starts is a decision the client takes with the number in front of them.
It also produces better decisions. Clients frequently decline changes once they see the price, and a change declined costs nothing. A change executed and priced afterwards costs whatever it costs.
Where the value genuinely cannot be established quickly, an estimate with a stated basis and a cap is better than proceeding blind. On prime residential, where the client is personally involved and making decisions continuously, this is the whole of cost control, as set out under prime residential.
Four, close it in the next valuation
An agreed change should appear in the following month's valuation and then be closed. Items carried as pending for months accumulate into the final account, which is where they become disputed.
A useful metric is the age of the oldest open item. Where that number grows every month, the process has stopped working regardless of how many registers exist.
The alternative is the position described in late design changes: a final account carrying eighteen months of unvalued work, argued about by people who were not there.
The programme half
Change has a time consequence as well as a cost one, and valuing the money without addressing the time is half the job. A variation instructed with no view on programme becomes an extension of time claim later.
The practical approach is to require a programme impact statement with any significant change, agreed at the same time as the value. Neither party benefits from leaving it open.
Where the sequence itself is changed without the quantities moving, that is still a change and it should be valued as one, even though no variation instruction may have been issued.
Why it fails on good projects
Change control usually breaks under programme pressure, when everybody is trying to keep the job moving and the paperwork is the thing that gets deferred.
That is precisely when it matters most, because programme pressure generates the most change and the least record. The projects that come through it are the ones where the habit is strong enough to survive being inconvenient.
The cost of maintaining it is a few hours a week. The cost of abandoning it is measured at the final account.
What this means for you
Track the age of open change items rather than the count. A rising number is the earliest reliable signal that the process has stopped working.
Then insist on valuation before execution as a rule rather than an aspiration. It is unpopular for about three weeks and then it becomes how the project runs. Our approach sits under cost control and variations.
Change piling up faster than anyone can value it?
Send the instruction log and the last valuation. We will tell you where the process has broken and what it will cost to leave it.