CVEN9702 Chap.3 Estimating Durations and Productivity
Estimating Durations and Productivity
A schedule duration combines work quantity, production rate, discrete setup or testing work and a calendar. The basis may come from measured work, practical experience, subcontractor advice, public approval periods or vendor fabrication data. Estimate detail should match project maturity.
Learning effects can reduce average labour across repeated units, but only when work, design and workforce conditions remain sufficiently stable.
What this chapter covers
- 01
Separate quantity and productivity from elapsed calendar time
- 02
Match estimate detail to the information available
- 03
Use subcontractor, authority and vendor durations transparently
- 04
Assign the correct calendar to each activity
- 05
Apply learning curves only to comparable repeated work
- 06
Record ranges, assumptions and triggers for revision
Worked example · free
Estimate a drainage installation duration
- +1Production work takes 480 divided by 32, which equals fifteen working days.
- +1Add one discrete working day for mobilisation and first-section testing.
- +1The total is sixteen working days, not sixteen calendar days.
- +1Assign the stated working calendar and record ground, access, crew and productivity assumptions.
Key terms
- Production rate
- The quantity of acceptable work completed per unit of working time under stated conditions.
- Working calendar
- The pattern of working and non-working periods used to convert workdays into elapsed dates.
- Estimate basis
- The quantities, rates, evidence, assumptions and date supporting a duration or cost estimate.
- Learning curve
- A model in which average effort per unit falls as cumulative production repeatedly doubles.
- External duration
- A waiting or delivery period supplied by an authority, vendor or subcontractor rather than calculated from internal productivity.
Estimating Durations and Productivity FAQ
Why are working days different from calendar days?
A working duration counts only periods when the assigned calendar permits work. Weekends, holidays, shutdowns and activity-specific calendars change elapsed dates. Concrete curing, normal labour and crash work may therefore use different calendars in the same project.
When should a learning curve be used?
Use it for repeated comparable operations with stable design, methods and workforce continuity. Disruption, redesign or changing crews can interrupt learning. Preserve the model basis and avoid treating one generic coefficient as a universal construction fact.
How should uncertainty appear in a duration estimate?
Record an expected value, plausible range and the assumption that drives the range. A deterministic duration is still needed for the baseline, but retaining uncertainty prepares the estimate for PERT, simulation and later review.
How can a duration estimate be validated early?
Compare the units and crew logic with a trial section, current supplier advice or measured output. Record the result before one unsupported rate is repeated across the programme.
Exam move
Build duration estimates from units. Start with quantity and a production rate carrying matching units, then add setup, inspection or curing work that does not belong inside the rate. Assign the work calendar only after the working duration is clear. Practise translating sixteen working days into dates on five-day, six-day and continuous calendars so calendar errors become obvious.
For external activities, locate the accountable duration from the current authority, vendor or subcontractor and record its date rather than relying on memory. Revise learning curves by calculating cumulative average effort after successive doublings and by distinguishing average from marginal unit effort. In written answers, name the conditions required for learning to occur.
Keep an estimate basis note beside every practice problem: scope, quantity, crew, rate, calendar, exclusions and revision trigger. That note is often more valuable than another decimal place. Build three estimates for the same work using measured output, a crew composition calculation and a benchmark rate. Reconcile the differences instead of averaging them blindly.
Check whether the quoted rate describes gross shift time, productive hours or completed work accepted after testing; those denominators produce different durations. Separate fixed work such as mobilisation, permits and commissioning from variable work proportional to quantity. Test the estimate with a units line so metres divided by metres per day visibly returns working days.
Then convert working days to dates with the nominated calendar, holidays and shift pattern. For repeated units, calculate learning only when design, method and team continuity justify it, and state what event would reset the curve. Create a low, central and high productivity case and identify which assumption drives the spread.
Compare that range with schedule float and contractual milestones rather than presenting false precision. When an authority or vendor controls an external lead time, record the source, retrieval date and follow-up trigger. End each practice answer by naming one validation action, such as a trial section, supplier confirmation or comparison with actual output from the first week.
This closes the estimating loop: an estimate begins as an explicit hypothesis, is scheduled with the correct calendar and is revised when project evidence becomes available.
