Compared with what would have happened anyway?
An improvement after a change does not reveal how much the change caused. A shop may become busier after replacing its sign, while nearby shops also become busier during a festival. The useful comparison is with a credible estimate of what would have happened without the replacement.
Go deeper
The missing comparison is the key
A shop replaces its sign, and the following week's visitor count rises from 100 to 130. It is tempting to credit the sign with all thirty additional visits.
But a festival may have increased foot traffic across the area. The shop might have attracted more visitors even with the old sign.
A counterfactual is the outcome that would have occurred without the change. We cannot observe both versions of the same shop at the same time, so we need a defensible comparison.
Similar unchanged shops, repeated measurements or a well-designed experiment can help estimate that missing outcome. Each method has limitations. A simple before-and-after difference includes the intervention and everything else that changed during the period.