If segmentation does not bring us a larger quantity of responses from a single offer, why bother doing it?
The answer is that you don’t do segmentation to increase the quantity of responses from a single offer. You use segmentation to optimise the effectiveness of a single offer. For example, by finding the 20% of the contacts who deliver 80% of the positive responses.
When we run a promotion to a 20% segment of the database, we make a conscious decision not to communicate with the other 80%. We therefore have an opportunity to present these people with a different offer – one that is more relevant to their needs.
To increase the total quantity of responses within any time period, you need to run multiple offers in parallel. And this means investing additional resources.
Let’s push these ideas about segmentation to their logical conclusion, so we can clearly see their limits. In an ideal world with a near-perfect customer database we would have:
- five clearly defined segments that don’t overlap, each representing about 20% of our database,
- five distinct offers that don’t overlap, customised for each of these segments.
And the results would be five sets of responses, each representing about 80% of a ‘normal’ response rate. The total quantity of responses from five promotions should be (5 x 80%) = 400% or four times the ‘normal’ response.
Let’s keep an eye firmly on reality. This best-case scenario tells us that, to get a four-fold increase in the quantity of results, we have to ramp up our promotional efforts by a factor of five. In other words, segmentation is subject to diminishing returns.