The Measurement Minute by Gary Angel

 

For shopper measurement, self-selection introduces massive biases in behavioral data. Different shopper populations behave differently not because of what we do or how the store works, but simply because they have different intent. When two stores perform differently, it may be because one store is better than the other. But just as or even more likely is that it’s because one store has a shopper population with higher buying intent. What all this means is that when using seemingly straightforward metrics like Conversion per Opportunity, we need a way to contextualize them with shopper intent.