This week I made a presentation to a large telecom client of ours in Istanbul about the virtues and how-to's of customer analytics and segmentation. As a normal part of the transition strategy to becoming more customer-centric, we usually suggest that a client should identify one or two segments, and place them into a pilot program of customer management. By running this pilot project over a number of months, you can work out the kinks and conflicts involved in managing customers, rather than just products and channels. You can also publicize your success within the company to help secure support for the overall transition. At the end of my talk, one of the senior executives asked me a question that got me thinking very carefully: Are there any particular criteria we should employ to choose which segments to put into a pilot program?
Now this is one of those questions that I ought to have answered a dozen times already, but I don't think I'd ever been asked, nor have Martha and I actually written anything on this topic. Off the cuff, really, I enumerated some criteria, but I'd be very interested if any of our readers could contribute to this discussion.
What I suggested was choosing a segment
- for which you have reasonably accurate and complete data,
- with a not-too-complicated set of channel or distribution conflicts,
- that already tends to buy from multiple business units, channels, or divisions,
- where you think you have a reasonable chance of creating significant value, and
- you can describe easily to non-statistical types.
I think the reasons for all my criteria are self-evident. But what did I miss? Any additional suggestions?