What is one more person on the gang worth?
Twenty-two site days regressed on crew size, late deliveries and rainfall. Crew size is worth 14.3 square meters a head with a tight interval, and the two things the site diary blames come back with confidence intervals that straddle zero.
Construction Intermediate Regression free
After you install, this is the model to open.
What Explains How Much a Crew Gets Done?
- In your spreadsheet, click the Sortia icon in the strip of icons down the right-hand edge. No strip? Click the arrow at the bottom-right to open it. You can also use Extensions, then Sortia, then Open Sortia.
- Click Start from a template and put that name in the search box.
- Pick the card with that name and click Load this template. It arrives on a new tab with real numbers already in it.
The answer
- One more person on the gang
- 14.3 m2 a day, with a 95% interval of 11.2 to 17.4
- The model holds
- R² 0.92 crew size does almost all of the explaining
- Late deliveries
- -1.09 p = 0.59, the interval straddles zero
- Rain
- -1.32 per mm, p = 0.11, the interval straddles zero too
Twenty-two site days of blockwork with three things the site already records. Click Run. R2 comes back 0.9217 and the standard error is 8.63 square meters, so this model explains most of the variation in a day's output and its typical miss is under nine square meters. The crew-size coefficient is 14.26 with a standard error of 1.48, a t of 9.65 and a 95% interval from 11.15 to 17.36: every extra person on the gang is worth between eleven and seventeen square meters a day, and that interval is the number to quote when somebody asks what a laborer buys you.
Now read the other two coefficients properly, because this is where regression reports get misused. Late deliveries come back at -1.09 with a p-value of 0.586 and an interval from -5.20 to +3.03. Rain comes back at -1.32 a millimeter with a p-value of 0.115 and an interval from -3.00 to +0.35. Both intervals contain zero, so on this month's evidence neither has a demonstrated effect on output.
That is not the same as proving they do nothing. The rain interval is mostly negative and the coefficient has the sign you would expect, so the honest reading is that twenty-two days is not enough to measure a rain effect this small, not that rain does not matter. Leave the correlation matrix on, which the prefill does, and the rain result gets more interesting rather than less.
Rain correlates -0.709 with output on its own, which looks like a strong effect, and -0.648 with crew size. The site already responds to rain by sending people home, so most of what rain does to output arrives through the crew-size column, and once crew size is in the model there is very little left for rain to explain. That is not the model being wrong, it is the model telling you that rain costs you people rather than costing you productivity, which is a different problem with a different fix.
Rerun with crew size dropped from the predictors and R2 falls from 0.922 to 0.517, which is the cleanest way to see how much of the answer that one column owns. The per-person block at the side is the sanity check on all of it: the best day produced 18.7 square meters a head and the worst produced 14.8, so a bigger gang lays more and does not lay more each.
What the model cannot tell you is what a bigger crew costs. To use your own diary, put daily output in the first column of the sheet, keep the predictors in the next contiguous columns, and widen both ranges to your rows.
The model
It arrives on a tab called Template: What Explains a Crew Day, carrying these columns:
- Blockwork laid (m2)
- Crew size (people)
- Late deliveries (count)
- Rain (mm)
- Date
with the model computed beside the data:
| Output per person, best day | calculated |
| Output per person, worst day | calculated |
| Days in this sample | 22 |
Once it is in your sheet
- The model arrives with real numbers in it and runs as it stands, so you can press the button first and understand it second.
- Change the numbers to yours. The sheet marks which cells are inputs and which hold formulas, and most labels carry a note explaining the row.
- Press the run button at the bottom of the panel. It is labeled for the tool you are in, and the result lands on its own tab, with a written reading of it beside the figures.
Never used Google Sheets? Start here goes the whole way, in seven steps, and assumes nothing.
Next question
- Are you paying for machine days you never use?Which Plant Do We Hire and Which Do We Own?
- Which tenders should you not price at all?Score the Sites Before You Bid
- Do rain days really cost you money?What Moves With Cost on This Job?
- Is one estimator quietly pricing higher than the other?Do Our Two Estimators Price the Same Job the Same Way?
- What will a tonne of steel cost you next month?Where Is This Material Price Going?
- How much contingency does the job need before the budget holds nine times in ten?What contingency gives 90% odds of staying under budget?
Every model like this one, and the method behind them: Statistics in Google Sheets.