In keenness to attribute success or failure to an intervention, we often forget

Every turnaround has a hero standing suspiciously close to the bounce. The worse the starting point, the better the photograph afterwards.

An extreme result often drifts towards average all by itself. The dreadful quarter is followed by a merely poor one. The spectacular campaign settles down. A customer selected precisely because usage collapsed improves after outreach, partly because the original month was unusually grim. Then the intervention claims the movement like a cat sitting on a parcel. This is wonderfully convenient for anyone whose programme began at the bottom of a chart.

Before-and-after charts love this trick. They begin at the most dramatic point, insert a manager, campaign or reform, and let two dots imply causality. Longer history tends to spoil the photograph. So does a comparison with similar customers who received nothing and improved anyway.

Ask what probably happens without the action. Use a comparison group where possible. Look at repeated measurements and choose the baseline before seeing which one makes the intervention heroic. Keep the ugly months on the page. The action may still deserve credit. It simply has to earn it after the ordinary bounce has been deducted.

Be particularly suspicious when the timing flatters whoever approved the spend. A sales slump, service spike or freakishly good launch already contains pressure to move.

Statistical gravity has no personal brand, no keynote and a hopeless LinkedIn profile. Humans keep accepting its awards.

Behavioural principles

Behavioural ideas at play in this post

Short, plain-English explanations of the principles behind this post, with links to related books and examples in the archive.