Construction Cost Overruns: How to Stop Margin Erosion Before Job Close

The job bid at $1.2M. It closes at $1.32M. The 10 percent overrun came from somewhere.

Ask the owner where. He can name two of the five places. The other three are where more than half of the gap actually lives.

This isn't a story about one bad job. It's a story about the pattern. Construction cost overruns are not random events. They're the visible symptom of five financial discipline gaps that compound across every open job, every month. If you can name them, you can catch them. If you can catch them, the variance from bid to bank starts to compress.

We work with $20M to $100M contractors trying to close that gap. After more than a dozen Civil CFO engagements, the five places cost overruns hide are consistent enough that we can predict them before opening the books.

Let's walk them.

What cost overruns actually mean for your business

A 10 percent overrun on one $1M job is annoying. A 10 percent overrun on every job for a year is your entire net margin.

That's the math owners miss. The overrun on any single job feels like a story (the weather, the sub, the design change). The aggregate of overruns across the portfolio is the explanation for why the company that's bidding at 22 percent gross is banking at 3 percent net.

Take a $30M contractor running 25 jobs in a year. If the average overrun is 8 percent, that's $2.4M of cost that wasn't in the bid. On a business banking $900K of net, the overrun aggregate is bigger than the entire net profit.

This is why cost overruns deserve attention even when no single job feels alarming. The killer isn't the outlier. It's the silent average.

The 5 places cost overruns hide

1. Productivity assumptions that don't hold

The estimating sheet assumes the crew can place 250 cubic yards of concrete per day. The crew actually places 190. That 25 percent productivity gap shows up in labor hours, equipment hours, and schedule slippage. Compound it across 30 days of pour and you have a major overrun no PM caused.

The fix isn't blame. It's feedback. Productivity assumptions need to be tested against actual job history every quarter. If your estimating team is using productivity numbers that haven't been validated against the last 24 months of closed jobs, your bids are structurally optimistic.

We worked with one $22M earthwork contractor whose estimating productivity assumption on cut-and-fill was 12 percent higher than what his own jobs had actually delivered over the prior three years. Twelve percent. On a job-by-job basis it looked normal. Across the portfolio it was a million dollars of structural margin loss per year.

2. Cost creep nobody catches monthly

This is profit fade, and it's the most common one.

The job is going. Month one looks fine. Month two, the labor budget is 8 percent ahead of plan. Month three, equipment hours are running long. Month four, a sub is behind and your crew is absorbing some of his work. Month six, a material price spike on remaining purchases.

Every one of those was visible the month it happened. None of them showed up at the monthly close because the close was a status update, not a forecast. The variance compounded silently until job close, when it showed up as "the job lost money."

The fix is the monthly WIP review running cost-to-complete on every open job, not just percent complete. More on that in a minute.

3. Unbilled change orders

The customer asks for extra scope. The super agrees. The crew does the work. The cost lands in your job costing. The revenue never does.

We've audited contractors at this size and found tens to hundreds of thousands of dollars of unbilled change order work in a single year. Some of it the owner knew about. Most of it he didn't.

This isn't an overrun in the traditional sense. The cost was incurred. The problem is the cost wasn't matched with the revenue it was supposed to capture. From a margin perspective, an unbilled change order looks identical to an overrun. The cost shows up. The revenue doesn't.

The fix is a written change order policy enforced in the field. No work outside scope without a signed authorization. The first 90 days are uncomfortable because field culture has to change. The next 90 are clarifying because the unbilled work either stops happening or starts getting captured.

4. Material price swings between bid and procurement

You bid the job in February. You buy the steel in May. The price moved 7 percent in those three months. Nobody adjusted the job margin forecast. The overrun shows up in your cost report as if it were random.

Material price volatility is significant. The ENR Construction Cost Index tracks it monthly and shows swings of 5 to 15 percent on key commodities in normal years. In abnormal years (2021, 2022) the swings hit 30 percent plus on some materials.

The fix is two-part. First, lock material pricing where you can (price holds, contracts, early procurement on long-lead items). Second, where you can't lock it, build a material escalation contingency into the bid that's calibrated to the volatility band. Free-floating material risk has to be either transferred or priced.

5. Field decisions made without cost data

A foreman makes a call in the field. Send three trucks instead of two. Run a sixth man on the crew for safety on this lift. Pour an extra load because it's there. Each decision is small. Each one costs money. None of them is visible in the cost report until weeks later.

This is the place cost overruns are hardest to attack, because the field decisions are usually right operationally. The problem isn't the decision. The problem is the visibility loop.

In high-performing crews, the foreman has access to current job cost data. He knows what's been committed, what's been spent, and where the budget stands today. That single visibility loop changes the field decision-making rhythm. We've watched contractors install daily cost coding by foreman and recover 1 to 2 points of margin from this dynamic alone.