Pricing the job was the easy part. The real test is three months in, when half the trades are on site, three things have changed since the tender, and you're trying to work out whether you're still inside the number you gave the client. Managing a construction budget on site is where most of the margin is actually won or lost, and most contractors find out they've lost it too late to do anything about it.
Why the budget slips after day one
A budget priced at tender and never looked at again is only accurate on the day it was written. Every variation, every changed spec, every item that turns out different on site chips away at it. If nobody's tracking those changes against the original breakdown, they all land at once at the final account, as one big unpleasant number instead of a series of small ones you could have managed.
It's rarely one disaster. It's a steady drip: a client's change of mind on the kitchen layout, a bit of extra excavation where the ground was soft, a week lost waiting for a window delivery. Each one looks too small to chase at the time. Together they're the difference between a job that paid and a job that didn't.
Managing a construction budget on site, line by line
The only way to know where you stand is to keep the budget live, checking actual cost against the original priced item, not the total. If the groundworks line was priced at a set figure and the ground's turned out different, you need to know that the day it happens, not the day you reconcile the account.
That only works if the original breakdown was measured properly in the first place, to NRM2, with every element on its own line. A single tender total gives you nothing to compare against except itself.
A simple routine that works on most small and medium jobs:
- Set the baseline. Turn the priced breakdown into your cost budget before the job starts, element by element.
- Code every cost. Each invoice, timesheet and hire charge gets allocated to the element it belongs to, not dumped into "general".
- Compare monthly at least. Actual cost to date against the value of work done for each line. Weekly on fast jobs.
- Log every change. Record each instruction or change the day it's given, with who asked for it, even if it seems minor.
- Forecast the finish. For each line, estimate the cost to complete. What matters is where it'll land, not just where it is now.
- Act on the gaps. A line running over is a question to answer this week, not a note for the final account.
Variations: price them while they're fresh
Variations are where most on-site budgets fall apart. The client asks for something extra on a Tuesday, it gets done on the Thursday, and nobody writes a price down until the final account, by which point everyone remembers it differently.
Your contract will usually say how changes are instructed and valued. Under JCT forms they're called variations and come through as instructions. Under NEC they're compensation events, with their own notification process and timescales, and the early warning process is there for flagging risks before they turn into cost. Whatever the form, the practical rule is the same: get the change in writing, price it quickly, and agree it before the money's spent if you possibly can.
When the original breakdown is measured, this gets much easier. You can price the variation against real numbers already in the contract: the same rates, the same quantities, adjusted for what's changed. It's hard for anyone to argue with a variation priced on the rates they accepted at tender.
A variation priced the week it happens is a conversation. A variation priced at final account is a dispute.
Watch the prelims as closely as the work
The line most likely to run away quietly is prelims. If the job's running late, you're paying for scaffold, welfare, supervision and site set-up for longer than you priced. That cost doesn't show up against any particular element, so it's easy to miss. Check time on site against the programme every month and ask whether the delay is yours, the client's, or something the contract lets you recover.
This is where a prelims build-up priced week by week earns its keep. If you know what a week on site costs you in scaffold, welfare and supervision, you can put a figure on a client-caused delay straight away. If prelims went in as a percentage, you're left trying to reverse-engineer a weekly cost months later, which rarely goes well.
What we set you up with
Send your drawings and we build that baseline before the job starts, item by item, so you've got a proper breakdown to run the job against, not just a total to hit. Everything's measured to NRM2 and priced on current UK rates, with prelims and provisional sums on their own lines. It's the same document you tender from and the same one you manage the job from, which means the numbers never have to be translated from one to the other.
If you want to go further into forecasting the final position on a live job, there's more in checking whether the job is still making the money you priced. The starting point is always the same, though: a client-ready estimate broken down far enough to manage from.
One question to ask every month
For each element, ask: if we finished this line today at the rate we're going, would it come in at or under what we priced? If the answer's no on more than one line, you've found your problem while there's still time to fix it. That single monthly question does more for managing a construction budget on site than any amount of software.