How to measure construction project performance
Measure construction project performance by comparing a short list of numbers against the bid and baseline every month: projected margin, cost variance, schedule variance, labor productivity, change order recovery, and safety. Track the trend, not just the snapshot.
To measure construction project performance, pick five to eight metrics, define each one the same way on every job, and compare them monthly against the original bid and baseline schedule. The most useful set for most contractors is projected gross margin against bid margin, cost variance by cost code, schedule variance against the baseline, labor productivity on self-performed work, change order recovery, and safety. Each number should come from a record (the job cost report, the schedule, timecards, the change order log), and each should be reviewed by trend. A job that is 2 points under bid margin and stable is in better shape than a job that is on budget today but has faded three months running.
Performance measurement is not about grading PMs. It is about seeing trouble while there is still time to fix it, and about feeding real numbers back into estimating so the next bid is better.
The metrics that matter
Projected gross margin vs bid margin
The single most important number. Projected margin equals revised contract amount minus projected final cost. Projected final cost is cost to date plus a field-based cost to complete. Compare it to the margin in the estimate. A shrinking projection is called margin fade.
Cost variance by cost code
Budget versus projected final cost for each cost code. This tells you where the money is going: labor, material, a specific sub, general conditions. Totals hide problems; cost codes show them.
Schedule variance
Forecast completion date against the baseline and contract dates, in days. Track critical path activities, not just the overall percent complete. A job can be 60 percent complete on paper and still be late if the critical work is behind.
Labor productivity
For self-performed work, compare actual hours per unit installed to the estimate. For example, hours per thousand square feet of drywall hung, or per linear foot of conduit. This is the earliest warning on labor-heavy trades. See labor productivity.
Change order recovery
How much of the changed work did you get paid for? Compare approved change orders to the cost of all extra work performed. Also track how long changes take from request to approval. Slow approvals often mean unpaid work.
Safety
Recordable incidents, near misses, and the incident rate. BLS publishes injury and illness incidence rates and offers an incidence rate calculator, which lets you compare your company to industry figures using the same method.
Earned value, when the job supports it
On larger jobs with a cost-loaded schedule, earned value compares the budgeted value of work actually completed to what was planned and what was spent. Many contractors use two ratios: a cost performance index (earned value divided by actual cost) and a schedule performance index (earned value divided by planned value). Values below 1.0 mean over cost or behind schedule. GAO's Cost Estimating and Assessment Guide covers earned value management in depth for anyone setting it up.
Leading vs lagging indicators
Margin and schedule variance are lagging indicators. By the time they move, the cause happened weeks ago. Pair them with a few leading indicators that move first.
- Labor hours against installed quantities, weekly. This moves before the margin does.
- Open RFIs past their needed-by date. Old RFIs turn into delays and rework.
- Extra work performed without a signed change order. Each one is margin at risk.
- Submittals not yet approved for long-lead items. These drive the schedule months out.
- Near misses reported. A crew that reports near misses is one that is paying attention; a sudden drop to zero can mean people stopped reporting.
Leading indicators are cheap to collect if your field records are in decent shape. They are also the numbers a PM can actually change this week, which is the whole point of measuring.
How to set up project performance measurement
- Define each metric in writing. What counts as cost to date, which date is the baseline, how percent complete is measured.
- Lock the baseline. Original budget by cost code and the baseline schedule are saved at the start and never edited. Changes go in as revisions, so variance stays honest.
- Get the field to forecast. Cost to complete comes from the PM and superintendent walking what is left, not from the budget minus cost to date.
- Review monthly, with a weekly labor check. Monthly for margin and schedule, weekly for labor hours. See how to run a weekly job cost review.
- Compare trend, not just snapshot. Chart projected margin month by month for each job.
- Close the loop with estimating. After closeout, compare final unit costs to the estimate and update your estimating database.
Worked example: month five on a $950,000 job
Say a self-performing concrete and site contractor is five months into a $950,000 job with a 14 percent bid margin.
| Metric | Bid or baseline | Current | Read |
|---|---|---|---|
| Projected gross margin | 14.0% | 11.2% | Fading, down from 12.6% last month |
| Labor hours, flatwork | 0.11 hr per sq ft | 0.13 hr per sq ft | Crews 18% slow |
| Schedule | Complete week 30 | Forecast week 32 | Two weeks late |
| Change order recovery | All extra work billed | $14,200 of $21,000 approved | $6,800 unpriced or pending |
| Recordable incidents | 0 | 0 | On target |
The story is clear. Margin is fading because flatwork crews are slower than bid and some extra work has not been converted to change orders. The actions: review crew mix and pour sizes on flatwork, and get the $6,800 priced and submitted this week. The productivity number also goes back to estimating, because 0.11 hours per square foot may have been optimistic for this kind of site.
(Productivity math: 0.13 divided by 0.11 is about 1.18, so crews are using about 18 percent more hours per square foot than bid.)
Mistakes to avoid
- Measuring cost to date against budget. That shows spending, not performance. You need a cost to complete forecast.
- Letting the baseline move. If the budget is edited every month, variance disappears and so does the warning.
- Too many metrics. Twenty KPIs means nobody watches any of them.
- Ignoring unpriced extra work. It looks like a cost overrun until it is billed, and becomes one if it never is.
- No feedback to estimating. The same mistakes get bid again.
Templates, guides, and tools
A job cost report template and a WIP report template hold the core numbers. Read how to read a construction WIP report for the company-wide view and construction KPIs every owner should track for company-level metrics. Status reporting is covered in how to write a construction project status report.
Smart Business Automator includes job costing and WIP views by project, with AI risk alerts that flag jobs trending off plan. See job costing and WIP automation.
Frequently asked questions
What are the most important construction project KPIs?
For most contractors: projected gross margin against bid margin, cost variance by cost code, schedule variance against baseline, labor productivity on self-performed work, change order recovery, and safety incidents.
How often should project performance be measured?
Monthly for margin, cost to complete, and schedule, ideally tied to the WIP close. Weekly for labor hours on self-performed work, since labor overruns grow fastest.
What is cost performance index in construction?
It is an earned value ratio: the budgeted value of work completed divided by the actual cost of that work. Below 1.0 means the work is costing more than planned.
How do you measure labor productivity in construction?
Divide actual labor hours by units installed (square feet, linear feet, fixtures) and compare to the hours per unit in your estimate. Track it weekly by cost code.
