The final account says the package cost more than the estimate. That is a result, but it is not yet a lesson. The scope may have increased, the work may have been resequenced, or the cost report may include items that sat elsewhere in the estimate. Replacing the old rate with final cost divided by original quantity would carry those differences into the next job without explanation.
Begin with a reconciliation at the same scope boundary. Match the estimate allowance, the delivered quantity and the recorded cost. Identify transfers between packages, variations and known omissions. If those records are incomplete, say so. A cautiously described observation is more useful than a confident benchmark built from incompatible data.
Separate quantity from performance
Consider an illustrative package allowed at 1,000 units and £20 per unit. The delivered quantity was 1,200 units at £22 per unit. The total movement is £6,400. Holding the original rate constant, the extra quantity accounts for £4,000. The remaining £2,400 is the rate effect applied to the delivered quantity. That split does not tell us why the rate changed, but it stops the whole overrun being described as poor productivity.
Illustrative bridge: 200 × £20 = £4,000; 1,200 × (£22 − £20) = £2,400. Together they reconcile £20,000 to £26,400.
The order of a variance bridge matters. This example values quantity change first and applies the rate change to actual quantity. State that convention and use it consistently. Then investigate the rate movement: labour hours, resource prices, waste, plant utilisation, subcontract scope and site constraints. Avoid allocating the same effect to more than one cause.
Record the conditions that made the result possible
A production figure needs a work description and operating conditions. Was access restricted? Were quantities repeated or fragmented? Did the crew have continuous work, or did permits and shutdown windows interrupt it? A wastewater pumping station tie-in during a short isolation window tells a different story from unrestricted pipe installation along an open corridor.
Include the source and confidence of the evidence. Directly recorded quantities and labour hours may support a stronger conclusion than a retrospective allocation from a broad cost code. Keep useful imperfect evidence, but do not conceal its limitations. The next estimator needs to decide whether the conditions are comparable before applying the lesson.
Give the lesson somewhere to go
Choose an action with an owner. A scope omission may belong in the package checklist. A recurring supplier qualification may belong in the enquiry template. A supported productivity finding may justify a library update, with its date, source, conditions and reviewer. Different findings require different changes; they should not all become a percentage added to the next estimate.
| Finding | Change to make |
|---|---|
| Scope boundary was missed | Amend the package review checklist |
| Supplier exclusions were unclear | Improve the enquiry and comparison template |
| Output assumption was unsupported | Add qualified production evidence to the library |
| Cost coding prevented comparison | Agree better capture with the delivery team |
Record who will make each change and how its use will be checked on the next relevant estimate.
Retain the previous library value and explain the new one. A rate library is easier to trust when a reviewer can see why it changed. If the evidence applies only to a particular work method or location, create a qualified entry rather than overwriting a general assumption. Revisit provisional lessons when better delivery data becomes available.
Check whether the action helped
At the next relevant estimate, ask whether the revised checklist or library entry was used and whether it improved the decision. This is a useful discussion between estimating and delivery, particularly where each team sees different parts of the result. The aim is a better method for the next job, with enough context that another person can apply it sensibly.