22nd November 2016
By Scott Logie, MD, Insight at REaD Group
Not just sex but age, household income, number of kids, car driven, property type, digital engagement, supermarkets shopped at and loads of other variables. As an analyst, or an ex-analyst who employs much cleverer people than I ever was, one of the joys of working at a data owner is how much data we have to play with.
I do firmly believe that there is no business problem that data can’t solve. Quite often the challenge is getting the right data to be able to solve the problem. This can lead to long lead times while research is carried out or additional data sourced. For most of the analysis we do we combine the data that REaD Group hold on every individual in the UK with the data our clients hold on their customers to develop outcomes to help solve our client’s challenges.
One of the questions I am asked most often is what analysis I would do if I had free reign and the time and data to be able to carry out any project. While there are lots of really good options, such propensity modelling, attrition analysis, lifetime value analysis, next best product or marketing mix modelling, my own personal choice would be a segmentation of the customer base.
Segmentation is sometimes sneered at a bit by analysts, and indeed by marketers. What will you do with it? What are the directly attributable outcomes? What benefits will it bring?
For me though, the amount of work that has to be done to get a good segmentation delivers benefits in itself, never mind the outcomes of the project. For example, as part of the initial work there are always profiles created; a need to understand who are active and lapsed customers and what the retention rates are; who is providing the most value and through what products; what channels customers are engaging through and what impact campaigns are having – that’s lots of projects rolled into one!
And that’s before you really get into the outcomes of the segmentation itself which can often help provide real insight into the customer base, in a way that even profiling can’t provide. The multi-dimensional aspect of segmentation means that sub-groups are often created that just wouldn’t be found otherwise. With a client I was working with recently we unearthed a group of well-off middle aged mums spending a lot of money which was a segment the client didn’t even know they had.
In addition, there are cultural elements to a good segmentation that you don’t always get from other insight projects. Once the initial segmentation is completed, getting the personas built, the naming of the segments agreed and the strategies on how to manage the customer groups developed takes much more than a team of analyst in a dark room.
This creates a real opportunity to open up the data to a wider group of people across the company who can get involved and really get under the skin of the customer groups created. In my experience, this additional knowledge can often help guide and develop the segments to such an extent that the data can both educate the client but also the other way around. I have seen many customer groups broken into two, or combined together to create other segments based on inherent business knowledge that the data would never find. This is often the most fun part of the project.
For many on-line, or fast growing businesses, there are just so many things going on, often around service and responding to client queries that actually getting to know who the customers are is a step too far. There is a real irony here as frequently deep customer engagement is what made them successful to begin with but as they have added millions of new customers that is lost. Segmentation can really help redress that balance.
Of course, it isn’t just new businesses. Many years ago I worked at Bank of Scotland, when it was a really well run, well managed small local bank. We were very lucky to have a great CRM database, a good insight team and a boss at the time who really bought into segmentation. We spent a lot of time breaking the base down into small, manageable groups, getting to know them in detail then creating contact strategies to deliver relevant campaigns. The work was justified by astounding response rates – up to 25% in some of the segments. Even a postcard to just say thanks to an older, long-time customer group brought new products opened by over 10% of them.
Maybe this is where my love for segmentation started as it showed that there are actionable outputs and business transformation can take place with deep customer understanding and clever strategic implementation.
So while a propensity model might give you directly attributable income, and help decide who to target for a campaign, for me a segmentation provides so much more. And now with the additional variables that can be added from the vast quantities of data that REaD Group holds, the joy of sex, age, income and the rest really does add an extra dimension to the segmentations we build for our clients.