Tracking Crew Productivity in Units, Not Just Dollars

The Limit of Dollar-Only Job Costing

Your job cost report says framing labor is $8,000 over budget. Three explanations fit that number equally well: the crew was slower than estimated, the estimate assumed a rate you never pay, or the scope grew.

Dollars alone cannot separate them. Only units can — hours against quantity installed.

The Measure That Matters

Productivity is a unit rate: labor hours per unit of work. Hours per square foot of drywall. Hours per linear foot of pipe. Hours per cubic yard placed. Hours per fixture set.

Once you have it, comparison becomes possible in a way dollars never allow:

• Estimated 0.08 hours per square foot, achieving 0.11 — running 38% over on productivity
• Same crew on the previous job achieved 0.075 — so this is job-specific, not a crew problem
• Across four jobs the average is 0.105 — so the estimate is wrong, not the crew

That third conclusion is worth more than any single job's variance, because it changes every bid you write from now on.

Capturing Quantities Without Killing the Foreman

The objection is always the same: the field will not record quantities. Usually true when you ask for too much.

Practical approach:

• Pick five to ten codes that carry most of your self-performed labor. Not fifty. Nobody needs productivity on miscellaneous clean-up.
• Use units the crew already thinks in. Panels hung, feet run, yards poured. Do not invent a measure that requires a calculation.
• Capture at the same time as hours. One extra field on the daily entry, not a separate report.
• Accept approximation. A foreman's estimate of 85% of a wall is fine. Precision matters less than consistency.

This works far better when the crew has seen what their coding produces. A foreman who has been shown his own numbers will record quantities. One who has never seen a report will not.

Reading the Result

Three patterns, three different responses.

One job off, others fine. Job-specific cause — access, weather, sequencing, an owner interfering. Look for a recoverable claim before assuming it was your fault.

One crew off across jobs. A supervision or skills issue, and a conversation rather than an accounting entry.

Everything off by a similar margin. Your estimating database is wrong. This is the most valuable finding and the most commonly missed, because each individual job looks like a small excusable overrun.

Feeding It Back Into Estimating

The point of productivity data is not the report. It is that next year's bids use rates you have actually achieved rather than rates someone typed in years ago.

Quarterly, take your top codes, compute achieved unit rates across completed jobs, and update the estimating assumptions. Keep the history — the trend matters more than any quarter.

Contractors who do this consistently end up bidding more accurately than competitors who are still guessing, which shows up as both a better hit rate and fewer jobs that lose money after award.

What It Costs to Start

One extra field on the time entry, eight cost codes, and a monthly review. That is the whole implementation. The hard part is sustaining it for two quarters before the data becomes useful enough that nobody wants to stop.

If your job costing captures dollars but not units, that gap is one of the things a diagnostic is built to surface.