Crew and Manhour Formulas
Last reviewed 9 September 20262,370 words11 min read
The arithmetic underneath every duration
🟩 In one line — Duration in working days equals manhours divided by crew hours per day, where manhours equals quantity times a unit rate in manhours per unit; every duration in a P6 file should be traceable to that one line, with the four numbers written down somewhere a reviewer can find them.
👤 Who this is for — Juniors building their first schedule from a BOQ, and mid-level planners who have been typing durations for years and need to be able to show the arithmetic when a client's reviewer asks. You should know what a calendar is in P6 and roughly what a norm or unit rate is.
First, let's be honest about why this page exists
Most durations in most schedules were never calculated. They were typed. The planner looked at the activity, thought "about three weeks," typed 15, and moved on. Sometimes that instinct is good. But it can't be shown to anyone, it can't be scaled when the quantity changes, and it can't be defended eighteen months later when the delay analyst asks why the baseline said 15 days for something that took 40.
The arithmetic is not difficult. What trips people up is the units — confusing output (m³ per day) with productivity (manhours per m³), applying an 8-hour norm to a 10-hour day without thinking, or loading a peak crew across a duration that starts with three men and a survey team. The formula is simple; the discipline is doing it every time and writing it down.
🟨 The standard — what "good" looks like
AACE RP 32R-04 (Determining Activity Durations) sets out that durations should be derived from quantities, productivity rates and the resources assigned, with the basis documented. It is the reference to quote when someone says durations are "judgement."
GAO Schedule Assessment Guide, Best Practice 4 requires durations to be realistic, based on the resources actually planned, and documented so they can be reproduced. Best Practice 3 requires resources to be assigned so that duration and resource are linked rather than independent guesses.
PMI Practice Standard for Scheduling treats duration estimating as an output of resource and productivity assumptions recorded in the schedule model, and expects the assumptions to be maintained alongside the model.
NEC3 and NEC4 Cl 31.2 require the programme to show the principal equipment and other resources the Contractor plans to use for each operation. That is only possible if durations were built from resources in the first place.
FIDIC 2017 Cl 8.3 requires a supporting report describing methods and resources by stage — again the crew and rate behind the bars.
DCMA Metric 8 (High Duration) flags activities over 44 working days. Long durations are usually the ones that were never calculated.
🟢 Rule to remember: every duration is a quantity divided by a defensible daily output — and if you can't write the four numbers on one line, you don't have a duration, you have a guess.
How it actually works
There are two ways to express labour productivity, and they are the inverse of each other. Get this straight before anything else.
| Term | Units | Example | Used for |
|---|---|---|---|
| Unit rate / norm | manhours per unit | 15 mh per tonne of rebar | Estimating manhours, tender norms, labour budgets |
| Output rate | units per day (or per manhour) | 1.2 t per fixer-day | Field planning, gang outputs, quick checks |
A 15 mh/t rate and a 10-hour day means one fixer places 0.67 t/day. Divide one into the other and you get the other. The mistake is multiplying when you should divide.
The four formulas
| Step | Formula | What it gives you |
|---|---|---|
| 1 | Manhours = Quantity × Unit rate (mh/unit) | Total labour effort |
| 2 | Crew hours per day = Crew size × Hours per shift | Daily labour capacity |
| 3 | Duration (working days) = Manhours ÷ Crew hours per day | The number that goes in P6 |
| 4 | Crew size = Manhours ÷ (Target duration × Hours per shift) | What you need to hit a date |
If your norm is a gang output (a formwork gang of 8 does 45 m² of wall shutter per day), skip the manhours and go straight to Duration = Quantity ÷ (Gang output × Number of gangs).
Rounding. Round up to whole shifts. A computed 13.4 days is 14 days. Nobody has ever achieved 0.4 of a day, and a P6 file full of 7.36-day durations tells a reviewer you've divided but haven't thought.
Direct versus total manhours. Most published norms are direct labour only — the crew at the workface. Supervision, plant operators, and general labour for housekeeping are typically added as a percentage (10–20%) in the budget. For duration, use direct crew hours; the indirects don't speed up the wall.
Peak crew is not average crew. Every activity ramps up and down. If you compute a 10-day duration with 20 men and the activity actually runs at 6, 12, 20, 20, 20, 20, 18, 14, 8, 4, the average is 14 — and you'll finish in 14 days, not 10. Either use an average crew of 70–80% of peak in the formula, or use the peak and accept the resulting duration is optimistic.
Workface capacity. Formula 4 will happily tell you 60 fixers finish a raft in 4 days. Before you believe it, ask whether 60 people can physically stand on it. Rules of thumb: rebar fixing runs out of room somewhere around 25–40 m² per fixer on a flat slab; blockwork is limited by lift height and scaffold, not bodies; welding is limited by joints available, not welders. Adding men beyond the workface limit adds cost and adds nothing to progress.
Getting it into P6. The calendar's hours per day matter. Set them at Enterprise → Calendars → Modify → Time Periods (or check Admin → Admin Preferences → Time Periods for whether calendars drive hours-per-day). A 10-hour calendar means a 14-day activity is stored as 140 hours; if a reviewer's default is 8 hours, they'll see 17.5 days. See hours-per-day-and-duration-display.
If you are resource loading, the cleanest route is Duration Type = Fixed Duration & Units: you type the duration from your calculation, load Budgeted Units (manhours) on the resource, and P6 works out Units/Time. Fixed Units/Time (you type the rate, P6 computes duration) works too but bites when someone changes the calendar later. See p6-duration-types-explained.
Illustrative order-of-magnitude rates (Gulf, experienced crews, normal conditions, direct labour; for sanity checks only — use the benchmark pages and your own as-built data for anything real):
| Work | Typical unit rate | Notes |
|---|---|---|
| Rebar fixing, slabs and rafts | 12–18 mh/t | Higher for columns and walls (18–25) |
| Slab formwork, table system | 0.5–0.9 mh/m² | Conventional ply 1.0–1.5 |
| Concrete placing by pump | 0.4–0.8 mh/m³ | Excludes curing and finishing |
| 200 mm blockwork | 1.0–1.5 mh/m² | Excludes lintels and chasing |
| Cable tray 300 mm, ceiling | 0.5–0.8 mh/m | Height and congestion push this up fast |
🟥 Where people go wrong
- Inverting the rate. 0.7 mh/m² of formwork becomes 0.7 m²/mh in someone's spreadsheet, and the duration comes out at roughly half of reality. Always write the units next to the number.
- Mixing the norm's shift with the site's shift. A norm of 15 mh/t doesn't care whether the day is 8 or 10 hours — manhours are manhours. But people then divide by 8 when the crew works 10, or vice versa. And if the day is 12 hours, remember that sustained long days reduce output per hour (see overtime-and-shift-arithmetic).
- Using peak crew across the whole duration. The histogram in the tender says 24 fixers. They'll be there for 40% of the activity. Durations built on peak crew are systematically 20–30% short.
- Ignoring the workface. The formula gives 4 days with 60 men; the slab holds 30. Reviewers with site experience catch this instantly and then distrust every other duration.
- Forgetting what the norm excludes. Curing, inspection holds, survey, crane waiting, and permit time are not in the fixing rate. Curing in particular is calendar days on a 7-day calendar, not working days — a 7-day cure on a 6-day calendar shows as 6 working days.
- Precision theatre. 7.36 days, 11.2 days, 4.8 days. Round up and move on. False precision is a signal to reviewers that nobody looked at the output.
- Doing the arithmetic and not keeping it. The calculation lived in a spreadsheet on the tender manager's laptop. Two years later the delay analyst has 2,400 durations and no basis for any of them. Put the four numbers in a UDF or the basis memorandum.
⚖️ When you're challenged
"Where did 14 days come from?" "180 tonnes of rebar at 15 manhours a tonne is 2,700 manhours. Twenty fixers on a 10-hour day is 200 manhours a day. That's 13.5, rounded to 14. The 15 mh/t is our as-built from the podium; the tender used 14."
"Can you do it in 10 days if I give you more men?" "Ten days needs 27 fixers, and the raft has room for that — about 45 m² each. So yes, physically. The constraint moves to rebar delivery and the crane, which we'd need to confirm. Give me until tomorrow and I'll tell you if the 27 can actually be fed."
"Your rate is slower than the subcontractor's tender rate." "Their tender rate was 12 mh/t. Our measured rate on the podium was 15, same subcontractor, same crews. I've used what we've actually achieved. If they hit 12 on the tower, we finish early — and I'd rather show early against a real number than late against a hopeful one."
"You told me 96 hours but P6 shows 12 days." "Same number. The activity's on the 8-hour office calendar instead of the 10-hour site calendar. I'll move it — it'll show 9.6, rounded to 10."
📄 Related pages
- Justify a Duration — how to present this arithmetic when the number is challenged
- Productivity Adjustment Factors — what multiplies the unit rate when conditions aren't normal
- Overtime and Shift Arithmetic — why 12 hours doesn't give 50% more output
- P6 Duration Types Explained — which duration type respects your typed number
- Hours per day and duration display — why your 14 days shows as 17.5 on someone else's screen
- Calculator crew duration — the formula in a fillable sheet
- Resource Loading and Histograms — when the manhours go into P6 and when they stay in Excel
- Benchmarks — the rate tables, with the conditions attached
✏️ Worked example
Raft foundation for a mid-rise commercial building in Riyadh. Quantities from the BOQ, rates from the contractor's as-built database from a similar raft the previous year. Six-day calendar, 10-hour shifts, autumn — no summer adjustment needed.
| Item | Quantity | Unit rate | Manhours | Crew | Crew mh/day | Computed | Adopted |
|---|---|---|---|---|---|---|---|
| Blinding and waterproofing | 1,300 m² | 0.25 mh/m² | 325 | 6 × 10h | 60 | 5.4 | 6 days |
| Edge formwork (1.5 m deep) | 420 m² | 0.8 mh/m² | 336 | 6 × 10h | 60 | 5.6 | 6 days |
| Rebar fixing (100 kg/m³) | 180 t | 15 mh/t | 2,700 | 20 × 10h | 200 | 13.5 | 14 days |
| Concrete pour, continuous | 1,800 m³ | — | — | 2 pumps @ 30 m³/h each | — | 30 hours | 2 days on a 24-hour calendar |
| Curing | — | — | — | — | — | 7 calendar days | 7 days on a 7-day calendar |
Notes on the decisions:
- Formwork runs alongside rebar (SS+3 from rebar start), not in series — the edge shutters go up while fixing proceeds inward.
- The pour is sized by pump output, not manhours. 1,800 m³ at 60 m³/h combined is 30 hours continuous; it's a night start on a 24-hour calendar because the concrete can't stop. Labour for the pour (12 men across two shifts) is loaded for cost but doesn't set the duration.
- Curing sits on a 7-day calendar so the 7 days are real days; on the 6-day calendar it would silently become 8 calendar days.
- Workface check on rebar: raft is roughly 1,200 m² of plan area; 20 fixers is 60 m² each — comfortable. Peak crew is used because rebar on a raft ramps quickly and the tail is short; I've accepted a small optimism there and noted it.
The PM asks for 10 days on rebar. Formula 4: 2,700 ÷ (10 × 10) = 27 fixers. Workface: 1,200 ÷ 27 = 44 m² per fixer — fine. Rebar delivery: 180 t over 10 days is 18 t/day, which is about one flatbed load a day at 20–25 t a trailer, into one gate that also serves the excavation. That is the real constraint, and it's a logistics answer, not a manhour answer. We agreed 12 days with 24 fixers and a second delivery window at night.
The whole basis went into a UDF on each activity (Enterprise → User Defined Fields → Activities: "Duration Basis", text) as a one-liner: 180 t × 15 mh/t ÷ (20 × 10) = 13.5 → 14d; rate from Project X podium as-built. Two years later, that line is what the delay analyst reads.
📖 References
- AACE International RP 32R-04, Determining Activity Durations
- AACE International RP 25R-03, Estimating Lost Labor Productivity in Construction Claims (for the productivity definitions)
- GAO Schedule Assessment Guide (GAO-16-89G), Best Practices 3 and 4
- PMI, Practice Standard for Scheduling, 2nd/3rd ed. — duration estimating and schedule model assumptions
- NEC3 ECC Cl 31.2; NEC4 ECC Cl 31.2
- FIDIC 2017 Conditions of Contract for Construction, Cl 8.3 (supporting report)
- DCMA 14-Point Assessment, Metric 8 (High Duration)
- Oracle Primavera P6 Professional User Guide — Calendars and Time Periods, Duration Types, User Defined Fields
Illustrative ranges, for sanity-checking only. Not a substitute for your own as-built data, and not for use as evidence in a contractual claim.
From the field
Experience from working planners. Unreviewed — read it as experience, not guidance.
Add what you know about crew and manhour formulas. What worked, what the consultant pushed back on, what you would do differently next time. A paragraph is plenty.
Contributors get their name and one link on the site — your own templates, course or consultancy. We take nothing and hold nothing.
Add your experience