In the world of risk management and modelling much revolves around the segmentation of data.
New customer segments, new attributes for segmentation, segment performance and reporting.
Indeed, much of the new development – ‘the edge’ – is also around this approach.
Obtaining new data, more detail to be able better segment, understand and hopefully model (future) predicted performance.
With the explosion of computing power, ready availability of data this has become ever more popular, with new, great tools on the market every day.
However we need to step back and ask the question, how much segmentation is too much segmentation?
- Replicating what is applicable in financial services, maybe a little too much segmentation in Telecommunications, and way too much in Utilities
- In a larger business, using 10, 100 or more segments may work well, but in smaller businesses the population in each segment maybe so small to be fairly meaningless statistically, and certainly hard to manage.
There is a cost to all this complexity.
It is certainly exciting, interesting and even fun to explore the data in ever more detail, however is it critical to manage, to ensure the level of detail and segmentation is relevant to the business and industry at hand.
Sometimes getting the basics right, with less segmentation can yield higher value.
Getting this balance right is in fact the tricky part..!