The Job Looked Profitable. Then We Found the Missing Hours.

A mid-size contractor once told me, “We’re profitable… we just never know which jobs did it.”

That sentence is the job-costing data gap in one line.

This is a real situation (names and numbers slightly anonymized), and it wasn’t caused by “bad people” or “no software.” It was caused by normal friction: crews trying to build things while the back office tries to price reality.

The moment it broke

End of month. The PM, ops lead, and finance sit down to review job profitability.

One project shows a healthy margin on paper. The site team feels good about it. The client is happy. Everyone expects a green tick.

Then finance asks the killer follow-up:

“Why do we have 200 hours in ‘General Labor’… but only 60 hours assigned to activities?”

Silence.

Because the profit number on the report wasn’t “wrong.” It was unsupported. Like a bridge with missing bolts.

The three gaps that created the illusion

1) Time went in… but not where it belonged

Hours were captured, but cost codes were messy (or optional, or late). A bunch of labor ended up in catch-all buckets because that’s what happens when the field is busy and the system is fussy.

This sounds small until you realize job costing is basically: estimate vs actuals by cost code. If the code is missing, the variance doesn’t scream. It whispers.

Industry job-costing guidance regularly flags coding errors, missing time entries, and late paperwork as common reasons job cost reports become unreliable. Source

2) Materials existed in real life… but not in the job

Deliveries arrived to a depot, then went to site “when needed.” Receipts landed in someone’s inbox. POs were created “later.” Some items got booked to the wrong job because the vendor invoice didn’t match the site reality.

So the project looked profitable this month… until the costs hit next month. That’s not profitability. That’s accounting time travel.

3) Change was happening… but wasn’t priced fast enough

Scope creep wasn’t dramatic. It was a hundred micro-decisions: “can you also…?” “while you’re there…?” “small tweak…”

Those changes were real work, but the pricing trail lagged behind the shovel trail. The job-cost report didn’t show the risk early, because the underlying cost and scope data wasn’t connected in time.

Construction project control research repeatedly highlights how changes and weak control processes contribute to overruns—problems that get amplified when cost data is late or fragmented. Source

What they did next (and why it worked)

They didn’t “transform the business.” They didn’t buy a shiny dashboard and pray.

They did something more boring and more effective: they treated job costing like a data supply chain.

They mapped the minimum dataset needed to trust a margin number:

  • every hour must have a job + cost code (and a default that isn’t “misc”)
  • every purchase must tie to a job before it becomes an invoice
  • every change must have a timestamp and a commercial status (priced / pending / rejected)

Then they tracked “gap rate,” not just cost—because reliable cost control depends on data being timely and consistent enough to act on.

Cost-estimating/control guidance also links weak cost practices to overruns and missed deadlines (the exact mechanisms vary by organization, but the relationship is well established). Source

And “getting the basics right” is unsexy—but it’s often where the biggest operational impact comes from. Source

The punchline

Two months later, the same contractor ran the review again.

This time, finance asked fewer “gotcha” questions. Not because projects were perfect—because the data trail was good enough to trust the conversation.

The biggest change wasn’t margin. It was confidence.

When job costing data gaps close, teams stop arguing about whose spreadsheet is right and start arguing about what to do next. That’s a better argument to have.


References: MASTT: Construction job costing guide · McKinsey: Construction productivity imperative · GAO: Cost estimating & assessment guide · CIOB WCS paper (project control/overruns)