Segments. But where are the oranges?

by | Sep 28, 2015

I started to write a piece about RFV (which I’ll now talk about another time). As I was writing I realised the purpose of my blogs is to get people thinking about data and sometimes people don’t think about it because they don’t know where to start. I live and breathe it so usually make the grand mistake of diving straight into something assuming everyone knows what I’m talking about.

So this time I’m going right back to basics: segmentation.

Imagine all of your data in a pile in front of you. It has some order or more precisely ‘consistency’ as each record is a customer and you have their name and address, details of their purchases and so on. But this doesn’t really help you because they are still in a big pile; like a mass of Lego pieces sitting on the table challenging you to make something.

As you look at the pile of chaos you start to see similarities between some of the customers. You decide to put everyone who’s made a recent purchase – let’s say in the last 12 months – in one pile and everyone else in another pile. You now have 2 groups, or segments.

Next up you can see that these two groups could be divided again, those who have purchased product A and those who have purchased product B. That makes 4 segments.

You can now start to look for more ways to split the groups; some people have made multiple purchases whilst others have only made one.

And your data looks like this:

Product A Product B
Single Multiple Single Multiple

0 – 12 months

Seg 01 Seg 02 Seg 03 Seg 04

12 months +

Seg 05 Seg 06 Seg 07 Seg 08


You’ve now have 8 segments. Every customer that sits in one segment has behaved in the same way as all the others in the same segment, during their lifetime with you. But how does it help?

Segmentation, however simple, helps you take the big chaotic pile of data and make sense of it. If you communicate one idea in one way to all your customers they will behave differently. If you split the chaos into segments there’s a good chance you will be able to see, not only, similarities in how the customers in one segment behaved in the past, but how they respond to current offers and communications, which means you should be able to predict, with some accuracy, what they’ll do in the future.

This is not magic. Neither is it particularly difficult. If you have no segmentation history you can keep the methodology really simple. Over time the methodology can and will evolve, changing from your naive first attempts to a more complex and brash entity. Chrysalis into butterfly!

Even with the most basic of data it should be possible to create some segments and define a few key groups of customers. There’s no reason not to – don’t be put off by thinking the simplicity you are able to conjure up has no value or is irrelevant. If you give it a go I can assure you, you’ll soon reap the rewards of the greater understanding of who’s keeping you in business. Let’s start looking into – and challenging – your data selections and start your transformation.


Share This