Big data is a d***

This will be an even more acceptable thought, after we’ve agreed that even ‘small data’ can be a d***.

Let’s look at a classic RFM (recency, frequency, monetary) segmentation. Any brand with a sizeable customer base can easily observe how most of their segments remain stable in terms of size and contribution. This may give them enough confidence to develop dedicated marketing tactics addressing these segments, as well as upwards migration (e.g. from lower to higher frequency).

But looking at aggregated data can be misleading. Individualized data often shows that customers do not remain stable inside an RFM segment, just waiting to migrate upwards as per plan.

Because most customers are promiscuous, and because shopping occasions vary significantly over time, the majority of customers actually fluctuate constantly between segments. These fluctuations might be invisible when observing aggregated data at segment level. However, individualized data often illuminates high rates of fluctuation, mostly without any discernible pattern, and (ironically) often counter to the marketing tactics applied at segment level.

The same is true for brand tracking, by the way: the same people can give different answers to the same questions, at different points in time. It’s complicated.

And we haven’t even talked about Big Data!