The short version
- Split labour into hours per unit and cost per hour. One is about the work, the other is about the market, and they move for completely different reasons.
- A wage rise then becomes a single edit rather than a re-price of every line.
- Height, repetition, access and sequence move productivity far more than crew quality does.
- Track actual hours on closed jobs against estimated hours. Without that feedback the library is just somebody else opinion.
What a composite rate actually is
A composite rate is a productivity assumption and a wage assumption multiplied together, then written down as one figure with both inputs thrown away. It works fine until either input changes, at which point you cannot adjust the rate because you no longer know what it was made of.
The alternative is one extra column. Hours per unit, cost per hour, and let the spreadsheet multiply. Every workbook we issue is built that way, so the labour column is a formula rather than a typed number.
| Composite | Split | |
|---|---|---|
| Item | Plasterboard to stud, 12.5mm | Plasterboard to stud, 12.5mm |
| Quantity | 2,400 m2 | 2,400 m2 |
| Labour input | $14.50 / m2 | 0.28 hrs/m2 at $51.80 / hr |
| Labour total | $34,800 | $34,800 |
| Wages rise 7% | Re-price every line by hand | Edit one cell |
| Ceiling goes to 4.2m | No mechanism | Hours to 0.36, cost follows |
| Compare against actuals | Not possible | 672 estimated hours against timesheets |
Why this became urgent
Labour has been the volatile input for three years running. Construction wages have been rising several points a year across most US markets, and the shortage of skilled trades is not a story about one region any more. If your estimating basis is a folder of composite rates, every one of them is quietly out of date and there is no way to tell which, or by how much.
With the split, a wage movement is a single edit to a rate table and every estimate referencing it moves together. That is the whole argument. It is not a theoretical improvement in rigour, it is the difference between a morning and a fortnight.
What actually moves the hours
Productivity assumptions are where estimates go wrong quietly, because a fifteen percent error in hours looks like nothing on a single line. These are the factors we adjust for, roughly in order of how much they matter.
| Factor | Typical effect on hours | Why |
|---|---|---|
| Working height above 3m | +15% to +40% | Access equipment, material handling, fatigue |
| Repetition, first unit to tenth | -20% to -35% | The learning curve is real and worth measuring |
| Restricted access or occupied building | +20% to +50% | Protection, out-of-hours working, small deliveries |
| Working out of sequence | +10% to +30% | Return visits, re-protection, trade stacking |
| Small quantities, under one day of work | +25% to +60% | Mobilisation dominates the task |
| Winter or exposed work | +10% to +25% | Weather, daylight, temporary heat |
| Complex geometry, curves and rakes | +15% to +45% | Setting out consumes more than the install |
Building the library without stopping work
- Start with the twenty line items that make up most of your turnover. Not all of them, the twenty.
- For each, write down the hours per unit you believe and the crew that does it. Your belief is a fine starting point.
- Price the next three bids with the split, alongside your usual method, and compare.
- When a job closes, put the actual hours next to the estimated hours for those twenty items.
- Adjust, and note why. A rate with a reason attached survives staff turnover.
Published productivity data is a reasonable starting point where you have no history, and we use it that way. But it describes an average crew on an average site, and the gap between that and your crew on this site is exactly the margin you are bidding.
What we hand over
Labour on our estimates is always a production rate times a wage rate, with both visible, plus the crew composition the rate assumes. If you disagree with a productivity figure you can change it and see the effect on the total immediately, which is the point. The estimate is a model of the job, not a verdict on it.
Written by the estimating team at QuantX Estimation. Since 2016 we have measured and priced construction work for contractors in the United States and United Kingdom, at a fixed fee agreed before we start.
