Catching Margin Drift on a Live 3PL Contract

Warehouse contracts almost never fail in one dramatic month. They erode — a shift that never quite came back down, a product mix that quietly got heavier, a pick rate that slipped after a WMS upgrade. Each one costs a fraction of a percent. By the time the annual review shows a margin two points below plan, the cause has been operating for eight months and the evidence trail has gone cold.

Why the annual review is the wrong instrument

A contract review compares a year of actuals against a year of budget and produces a variance. That variance is a single number describing twelve months of mixed causes. It tells you the margin moved. It does not tell you when it started, what changed, or whether it is still happening.

Worse, it arrives at the point where you can do least about it. Twelve months of under-recovery is money already gone. Three months of it, spotted in month three, is a conversation you can still have — about scope, about volumes, about a rate that no longer reflects the work.

What drift actually is

"The contract is losing money" is not a diagnosis. There are four distinct things that move a warehouse contract's margin, and they call for completely different responses:

Only the last of these is somebody else's fault. The other three are operational facts that your pricing needs to know about.

Measure against the plan, not the forecast

Here is the mistake that makes most drift monitoring useless: comparing actual cost to budget cost. That comparison is dominated by volume. If throughput ran 12% under forecast, labour cost will be under budget too, and the report says everything is fine — while the contract quietly under-recovers its fixed overhead.

The measurement that means something is:

This strips volume out of the picture entirely and leaves you looking at the thing you can act on: whether the work is costing what you said it would cost, per unit of work actually done.

Persistence and materiality: the two filters

A monitoring system that flags every monthly wobble gets switched off within a fortnight. Two filters make the difference between a signal and noise:

• Persistence. One bad month is a stocktake, a public holiday, a system cutover. A gap that holds for several consecutive months is a change in the operation. Require the pattern, not the point.
• Materiality. A 3% drift on a small activity is a rounding error; the same 3% on the main picking line is a car. Filter on the dollars at stake per year, not the percentage.

Set sensibly, these two filters mean a contract raises perhaps two or three findings a year — and each one is worth reading.

A finding needs an owner, not a dashboard

Most variance reporting fails at the last step. The number is correct, it is visible, and nobody does anything, because a dashboard belongs to everyone and therefore to no-one.

A drift finding should behave like a piece of work: assigned to a named person, carrying the month-by-month evidence that produced it, and closeable in one of two ways — actioned, or dismissed with a stated reason. The reason matters. "We knew, the customer added a new SKU range in March and we are re-rating at renewal" is a perfectly good answer, and recording it stops the same finding resurfacing every month while preserving the fact that it was considered.

What to do when you find it

Four responses, in rough order of preference:

1. Fix the operation. If productivity drifted because of a layout change or a training gap, the cheapest answer is to get the rate back. Drift detection tells you which activity to look at.
2. Re-scope. If the customer's work genuinely changed — new channel, new SKU profile, extra service — that is a variation conversation, and it is far easier to have in month four with six months of evidence than in month eleven.
3. Re-rate at renewal. Some drift is permanent and legitimate. Feeding the measured rate back into your site's own productivity assumptions means the next term is priced on what the site actually does.
4. Accept and record. Sometimes the answer is "this is a strategic account and we are carrying it". Fine — but it should be a decision on the record, not an accident discovered at year end.

The compounding benefit

Contracts monitored this way do more than protect their own margin. Every persistent drift finding is a data point about how your operation really performs, and those data points are exactly what should be pricing your next tender. An operator who measures drift is quietly building the most valuable asset in the business: a productivity model grounded in evidence rather than industry averages.

The principle

Margin does not leak because nobody cares. It leaks because the measurement arrives too late, is dominated by volume, and lands nowhere in particular. Compare actuals to the plan at the actual volumes, require persistence and materiality before you speak, and give every finding an owner. Do that and the annual review becomes a confirmation rather than a surprise.