The Real Reason Fixed-Price Engineering Jobs Keep Coming In Under Margin

The director of a UK fabrication business told me he would need to pull it all together to tell me what his live jobs were making. He had three jobs running. The orders had been placed. Work was progressing. He was running the business.

He did not know what those jobs were making.

Not because he lacked intelligence or experience. Because nobody had assembled the three numbers that would tell him. Materials spend against budget. Hours burned against the estimate. Third-party costs against the allowance. Those three lines existed somewhere in invoices, timesheets, and emails. But without a system to pull them together weekly, the only moment they arrived in the same place was the final account.

By then, the margin had usually been gone for months.

How Margin Leaves Without a Decision Being Made

Fixed-price engineering jobs lose margin in a particular way. There is rarely a single catastrophic decision. There are dozens of small ones. A scope addition gets absorbed without a variation being raised. Engineering hours run long and get written off internally. A bought-out item arrives late and triggers rework that nobody charges back.

Each event is survivable on its own. The accumulation of them is not.

On a £200,000 conveyor or fabrication job, 5% margin drift is £10,000 gone. On a £500,000 job it is £25,000. Neither figure announces itself. Both build across a project lifecycle without triggering a visible alarm, provided nobody is looking at the trajectory in real time.

The problem is not that the losses are invisible in principle. It is that most bespoke engineering businesses have no mechanism to make them visible in practice until the job closes.

What Pulling It Together Actually Looks Like

For a fixed-price job with three to twelve people working on it, the margin position can be tracked with three cost lines updated once a week.

First: materials. You placed the purchase orders. You know what has been committed and invoiced. Compare it to the materials budget in the original estimate. You are either tracking, over, or under. That is the first number.

Second: labour hours. Count the productive hours your team put into that job this week and multiply by your cost rate. Subtract from the hours remaining in the budget. If you are at 60% of your allowed hours and 70% of the programme has passed, the trajectory is probably right. If hours burned are at 80% of budget at the same point, you have a problem that is still recoverable, but only if you can see it now rather than at completion.

Third: third-party and logistics costs. You arranged them. The invoices exist or are pending. Add them to the running total.

Those three lines, updated weekly, give you a commercial position on the job. Not a full earned value analysis. Not perfect. But enough accuracy to know whether the trajectory is right and whether something needs to be challenged before the window to challenge it closes.

The argument that small bespoke jobs do not warrant a tracking process is wrong. On fixed-price work, the consequence of invisible drift is the same regardless of job value. It is a smaller number, which takes longer to become a problem. That means it becomes harder to see until it is too late to act.

What the Absence of Visibility Costs

On a bespoke automotive conveyor contracted at £1.2 million, the project manager believed he had contingency left in the estimate. He did not. The contingency had been distributed invisibly across multiple cost lines during estimating and trimmed during commercial negotiation. It was never tracked as a named figure with a running balance.

When a scope addition came in from site, he absorbed it on the basis of a buffer that no longer existed. The absorbed cost was between £26,000 and £28,000. Without a mechanism to make the commercial position visible before that decision was made, and without a variation record to support recovery, the job continued drifting.

By the time a commercial review was undertaken, the contract was heading for a loss. Recovery required excavating three years of verbal agreements and work delivered but never formalised or charged. The final account closed at £1.5 million against an original contract value of £1.2 million, with £300,000 in unclaimed variations recovered. It closed at 8% profit instead of at a loss.

None of that cost a pound of additional engineering resource. It cost margin that should never have been surrendered.

The External Environment Has Made This More Urgent

UK manufacturers experienced their largest monthly cost increase since 1992 in early 2026, driven by energy prices and rising steel costs. Structural sections are currently trading at £1,100 to £1,200 per tonne ex-works, with industry forecasts indicating increases of £80 to £200 per tonne across 2026 as import quota cuts and tariff reforms take effect from July. For any fixed-price project where steel was estimated before mid-2025 and procurement has not been locked in, the gap between the estimate and the actual cost is already widening.

Without a tracking mechanism, that gap becomes visible only at final account.

The conditions for silent margin erosion have worsened materially. The case for real-time commercial visibility has not been stronger.

The Minimum Standard

The question for any bespoke engineering business running fixed-price work is not whether the tracking system is sophisticated. It is whether you can answer, within five minutes, what each live job is making right now.

If the honest answer is that someone would need to pull it all together first, the visibility is not there. And the absence of visibility is not a neutral condition. It is the condition in which cost overruns accumulate before anyone decides to allow them.