Every managing director I sit down with can tell me last year’s turnover. Most can tell me the profit. Then I ask which jobs the profit came from, and the room goes quiet.
Someone says the school job did well. Someone else pulls a face, because they remember the school job differently. The honest answer, in most mid-market firms, is that nobody knows. Not because nobody cares — because nothing the firm runs can actually answer the question.
The accounts package can’t. Sage or Xero will tell you, reliably, whether the company as a whole made money. Ask it about one job and it goes vague, because that isn’t what it’s for. It sees invoices in and invoices out. It doesn’t see jobs.
Where job cost data actually sits
The cost of a job is scattered across half the business. Materials invoices arrive weeks after the lorry did. Labour sits in timesheets, or in a site manager’s memory of who was where. Subcontractors bill through their own applications on their own rhythm. Plant hire lands a month in arrears. By the time all of it has trickled into the accounts, the week it describes is long gone.
And that’s only the cost side. The other half of the question is what the work done so far is actually worth — how much of the contract you’ve genuinely earned by this point. On most jobs that figure only gets worked out once a month, when someone puts together the application for payment: the document you send the client saying here’s what we’ve built so far, and here’s what you owe us for it. Between applications, the number lives in somebody’s head.
The monthly job costing report
Quantity surveyors — the people whose trade is measuring and valuing construction work — have a name for the discipline that fixes this. It’s called cost value reconciliation, or CVR, and stripped of the jargon it’s one question asked monthly of every live job: what has this job cost us so far, and what is the work worth so far? The gap between those two numbers is your margin. Not the margin you priced. The margin you’re actually getting, while the job is still running and something can still be done about it.
The big contractors treat this as routine, like payroll. In the mid-market it is usually nobody’s job. Which means the first honest number most firms see is the final account — the closing bill for the whole project, settled months after the work finished.
That’s the problem in one line. The decisions that could have saved the margin — querying a subcontractor’s account, chasing the unpaid extras, pulling two men off a job that’s quietly drowning — needed making in November. You find out in March.
You cannot steer a job by its post-mortem.
There’s a second cost, and it’s sneakier. The rates that lost you money on the last job get typed into the next tender, because nothing ever flagged them as the culprit. A job that bleeds and is never diagnosed doesn’t just cost you once. It teaches your estimating to do it again.
This is not an industry with room for that. The most recent league table of the UK’s hundred biggest contractors put their average pre-tax margin at 2.4 per cent, with getting on for half of them under two. Mid-market margins are rarely more comfortable. At those levels, one job losing money quietly can wipe out three that did well — and you won’t know which was which.
Why job costing spreadsheets break down
Plenty of firms have tried to fix this with a job-costing spreadsheet, and I’ve read a lot of them. They fail the same way every time. Someone rekeys figures out of the accounts once a month, so the spreadsheet is always three weeks stale. And it only knows about invoices that have arrived — which misses the most important number of all.
That number is committed cost: money you’ve spent that nobody has billed you for yet. The steel you’ve ordered but not been invoiced for. The subbie who’s three weeks into a package and hasn’t applied yet. An order placed is money gone, whatever the accounts say. A spreadsheet fed only from the ledger leaves all of it out, so every job looks healthier than it is. False comfort is worse than no report at all.
What the fix actually looks like
Not a ten-module construction platform with a dashboard for everything and a licence fee to match. The tool this problem needs is small. It reads the accounts system you already run, so nothing is rekeyed. It picks up timesheets and orders placed, so committed cost is in the picture. And it shows one screen: every live job, cost so far, value so far, and which direction the gap has moved since last month. That’s the whole thing. When a job starts drifting, you see the drift in week two — not at the final account.
Where this doesn’t apply
If you run two or three jobs at a time and walk each of them every week, your head is the CVR, and it’s probably a good one. Don’t buy software to replace it. The same goes if you already employ a QS who produces a monthly CVR that actually gets read at a monthly meeting — that process is worth more than any tool, and nobody should sell you a replacement for something that works. And if your work is small, fixed-price and quick — jobs measured in weeks, not months — your margin was mostly decided the day you priced it. Your problem is quoting, which is a different conversation.
But if you’ve ever been surprised by a final account — the wrong way — then somewhere in your live jobs right now there’s a number moving against you, and nothing in the business is set up to show it to you. Finding that number is the first thing I look for when I audit a firm. If you’d like to know which of your jobs is paying for the others, get in touch.
