The factor matters more than the model

Two teams, one data set, a nine per cent gap

A client asked two firms to calculate the same Scope 2 figure from the same meter exports, as a check. The answers differed by nine per cent. Both arithmetics were correct. The models were, for practical purposes, identical.

The difference was a single emission factor. One team had used the current published set; the other was still carrying a 2019 grid factor in a lookup table that nobody had been asked to maintain. Nine per cent of a corporate electricity footprint, sitting in a cell no one had looked at in four years.

Where factors go stale

Rarely in the places people audit. The model gets reviewed every year because the model is the part everybody can see; the lookup table feeding it sits in a different file, belongs to nobody in particular, and has not been opened since the person who built it moved on.

• Grid intensity factors, which change every year and in some markets substantially
• Fuel factors carried forward from a previous consultant's spreadsheet
• Spend-based factors applied to a category that has since been measured properly
• Factor sets with no version recorded at all, which cannot be rebuilt later

What a factor register is for

Every figure carries its factor, the source it came from and the date that source was published. It is a dull document and it answers the only question that mattered in the example above — which, incidentally, took eleven minutes to settle once both registers were put side by side.

The factor set is versioned, so last year's number can be rebuilt exactly. Without that, a comparative is just a number that resembles the old one.

And when the factor changes

The prior year has to be rebuilt on the same basis or labelled as not comparable. Updating this year's factor while leaving last year's in place produces an improvement that nobody achieved, and an assurer will find it by doing exactly what the client above did.