Last week, I saw a presentation by Newcastle upon Tyne Hospitals Trust on their supply chain carbon footprint (usually the bulk of ‘Scope 3’ or indirect emissions). Like many organisations, they had been using ‘spend-based’ emissions factors for the purchased goods and services. These factors give you an average kgCO2e/£ spent on cardboard boxes/surgical scrubs/whatever, so you simply multiply the factor by your spend and, hey, presto, your answer. Simple!

The Trust is now moving from spend data to actual data from suppliers and what was glaringly obvious in the presentation was the actual data was about twice the spend-based data. Now while there is always going to be significant uncertainty in supply chain data due to availability, errors, time lags, sudden changes etc, that’s quite a margin of error.

But there are deeper problems with using spend data. The only way you can cut a spend-based carbon footprint is to spend less money. Inflation will make a static footprint appear to rise until the spend factors are updated. If you swap from an average supplier to a low carbon supplier, the carbon benefits of that change will not be reflected at all in your footprint. What’s worse, as Anindya Chatterjee of SmartCarbon pointed out when we discussed this over coffee, if you decide to pay a premium for a zero carbon product or service, your spend-based footprint will go up, not down. Which is bonkers!

Many of my colleagues in the sector tell me they still think spend-based factors are useful for a snapshot in time, but I see red flags waving all over the place. The biggest one is in presenting data to a non-specialist but important audience, eg the board of your organisation. You will constantly have to caveat the conclusions you are drawing from the data. That isn’t a great way to exude confidence in what you are presenting and is fraught with scope for misunderstanding.

The role I can see for spend-based data is to perform an initial screening exercise to separate significant suppliers from the rest – and then you can start gathering actual data from the former. In most cases the 80:20 rule applies, so at first you can get away with using spend data for the 80% of suppliers who contribute 20% of the footprint on the grounds that a ballpark figure is more accurate than zero, and using spend data here reduces your data burden, avoiding paralysis by analysis. But your significant suppliers’ contributions must be measured on actual data for your footprint to be not only meaningful, but in any way useful for decision-making and monitoring progress.

As you progress towards Net Zero, tackling those elephants in the room, you will have to start replacing the rest of the spend data over time, either by lowering your screening threshold or to reflect smaller suppliers who have made significant progress. Otherwise you may never know whether you have hit Net Zero, even if you have!

The broader point here is that many systems simply don’t measure the right things in the right way for Sustainability, and can produce misleading results. For example, one of my clients had an investment appraisal process that didn’t factor in carbon compliance costs, so it was no wonder that low carbon options were often passed over (now fixed). If a system isn’t fit for purpose – ie you find yourself having to explain away anomalous results or listing caveats – change it. This job is hard enough without trying to run up the down escalator.