top of page

MEP Crew-Balance Estimates: Why Total Labour-Hours Do Not Predict Daily Output

Oct 24
5 min read
Three workstations. One bottleneck. A worked planning model.

Original illustrative engineering schematic. The worked calculation below uses explicit example inputs.


An estimate can contain the correct labour-hours per assembly and still promise an impossible completion date. The missing link is often the sequence: one station cannot start until another has released its work, and the slowest repeating station limits output. Dividing total labour content by the number of available people assumes that work can be redistributed freely. Real access, tooling, qualifications and inspection boundaries may prevent that redistribution.


This example concerns a hypothetical bench preparation process for identical non-safety-critical MEP accessory kits. It is a planning calculation, not a prescribed installation method or labour standard. All activity times are illustrative measured-equivalent inputs. The actual method, workforce competence, handling arrangements and quality checks require project review before a production commitment.


1. Separate work content from the release interval


Labour content is the sum of hands-on person-minutes needed for one acceptable kit. Cycle time describes how often a station can repeat its complete task. Lead time is the elapsed time from a kit entering the first step until it leaves the final step. These quantities can have different values even when the same minute is their unit. The Lean Enterprise Institute distinguishes measured process cycle time and operator cycle time; its operator-balance approach makes unequal station workloads visible.


Assume three dedicated stations with one operator at each: A sorts and verifies components in six minutes; B assembles the kit in ten minutes; C labels and records it in four minutes. Each station processes one kit at a time. Transfers take negligible time, no rework occurs, and transfers are paced so that B always has the next kit ready. Station A may wait between kits rather than accumulate unlimited work in progress. There are no shared tools or other constraints in this simplified model.


Labour content per kit = 6 + 10 + 4 = 20 person-minutes.


Steady-state release interval = max(6, 10, 4) = 10 minutes per kit.


Steady-state capacity = 60 ÷ 10 = 6 kits/hour.


First-kit lead time = 6 + 10 + 4 = 20 minutes.


A common shortcut gives three people × 60 minutes ÷ 20 person-minutes = nine kits per hour. That is a theoretical perfectly balanced labour capacity, not the capacity of this allocation. Station B can process only six kits per hour. Extra idle time at A and C does not automatically become useful assembly time at B.


Compare station bars 6/10/4 minutes with 8/8/4 minutes. Show 36 kits: 370 minutes before, 300 minutes after; both contain 12 touch person-hours. Illustrative deterministic model, not a promised productivity result.

Original illustrative estimating graphic; all values are hypothetical worked-example inputs.


2. Include the first kit when estimating a finite batch


In a filled production line, consecutive completions are ten minutes apart. At the beginning of an empty line, however, the first completion takes twenty minutes. For a batch of N identical kits under these assumptions, the last completion occurs after the first-kit lead time plus N minus one bottleneck cycles. This is not the same as multiplying every kit by the full lead time, because several kits are being processed concurrently.


For 36 kits starting with an empty line:


Elapsed processing time = 20 + (36 − 1) × 10 = 370 minutes.


Touch labour = 36 × 20 = 720 person-minutes = 12 person-hours.


Three operators assigned for all 370 minutes = 1,110 person-minutes = 18.5 person-hours.


Difference = 390 person-minutes = 6.5 person-hours.


The difference is available station time not spent on the listed kit tasks. It is not automatically waste, a saving opportunity or a recoverable labour allowance. Operators may perform other authorised tasks, replenish components or complete necessary checks. If those activities occupy the same people or tools, put them into the model rather than counting their time twice.


3. Test a proposed rebalance explicitly


Suppose a reviewed method allows a two-minute preparation element to move from B to A without changing the total labour content. New station times are eight, eight and four minutes. This transfer is a hypothetical planning alternative, not permission to remove inspection, split an indivisible activity or assign work to an unqualified person. The first-kit lead time remains twenty minutes, but the repeating bottleneck interval becomes eight.


Rebalanced station times = 8, 8 and 4 minutes.


Steady-state capacity = 60 ÷ 8 = 7.5 kits/hour.


Thirty-six-kit elapsed time = 20 + 35 × 8 = 300 minutes.


Total touch labour remains 720 person-minutes.


With three operators present throughout, assigned time = 900 person-minutes = 15 person-hours.


The example predicts a shorter elapsed batch duration, not verified project improvement. Do not claim that reducing the calculated assigned time from 18.5 to 15 hours reduces payroll or project cost by the same amount. The crew may remain committed for a full shift, and other scope can absorb the released time. Test the method on representative work before changing the estimate.


4. Check the available production window


Assume the programme offers 360 net production minutes after separately identified breaks, briefings and planned interruptions. With the original allocation, the maximum completed batch is one plus floor((360 − 20)/10), or 35 kits. The rebalanced allocation yields one plus floor((360 − 20)/8), or 43 kits. Floor means rounding down; an unfinished kit is not a completed deliverable. Neither result includes defects or unplanned downtime.


  • Keep approved fixed setup and closeout time outside the repeating cycle unless the observation already includes them.

  • Use a distribution of observed times when variation is significant; averages alone can hide queues and interrupted work.

  • Model inspection availability, component replenishment and shared equipment as constraints, not unlimited support.

  • Separate accepted units from starts. Rework uses capacity and may revisit the bottleneck.

  • Record who approved each proposed work transfer and retain the original estimate as the comparison baseline.


5. Turn the calculation into an estimating decision


Use the result to challenge an output assumption, not to dictate a crew size. A useful estimate records the required quantity, station tasks, labour content, bottleneck interval, initial lead time and available window separately. If the required delivery exceeds calculated capacity, investigate a reviewed method change, another independently resourced line or a different sequence. Adding people without a feasible task allocation is not a capacity calculation.


The MEP Estimators Handbook is a broader estimating reference to use alongside a project-specific resource plan. Keep this worked model attached to the assumptions register so that the next revision can distinguish a changed quantity from a changed method or production window.



Continue the calculation review



Sources and calculation basis


Primary references checked on 24 October 2026. The worked examples and diagrams are original illustrative calculations, not manufacturer selections or project approvals. Apply the stated limitations and the requirements governing the actual project.




Comments

Rated 0 out of 5 stars.
No ratings yet

Add a rating
bottom of page