Segmentation - The Forgotten Hero

These days buzzwords seem to cloud the data and insights marketplace. Everyone seems to be talking about the future of predictive and machine learning algorithms. In the maze of buzzwords and sky high ROIs however, there is one very important aspect of analysis data and databases that is being side lined – segmentation.

Why Segment?

Segmentation, if done properly, allows you to find groups of customers within your database with similar behaviours. These groups are then easy to target with specific messaging tailored to their activities, making your life a whole lot simpler!

Let’s consider a scenario in which you want to send out a discount offer to your customers. If you have a group of customers who are happy to purchase without a discount, but sent the offer to everyone in your database, you would literally just be giving money away. Sending the same offer to everyone in your database could also potentially alienate your most loyal customers, they would realise there was no benefit being loyal to your brand as they get treated the same as everyone else.

What sort of segments and techniques should I use to split my database?

There are a number of methods you could use to segment your database:

• RFM Segments – Segment customers based on past purchasing history. RFM analysis groups individuals based on the Recency of their purchase, the Frequency of their transactions and the Monetary value of all transactions.
• Cluster Analysis – This will allow you to identify multiple groups of customers whose behaviour is similar in many ways. Any number of factors, dictated by business rules, can be used to group customers.
• Engagement Segments – You could classify your customers based on their interaction with your marketing communications. You can use these segments to tailor communication frequency and content type.
• Drive Time Segments – You can split your customers by their drive time or distance to your store/another point of interest. If you were to send store specific emails, you could then send these just to customers who live within a specific radius.

How can I use segmentation in everyday campaigns?

The simplest of segmentation allows you to tailor your communications and send to only relevant individuals within your database.

Utilising simple RFM segmentation could now transform your communications to your loyal customers, allowing you to identify the group of individuals who are most loyal to you based on their past purchasing activity.

Also consider splitting your database by the number of days since an individual has interacted with an email communication from you. If there are people who have not opened an email in three years, the simple engagement segments would allow you to exclude them in the future, in turn improving ROI.

What are the other benefits of segmentation?

Well firstly, segments are relatively easy to create and implement within your database. RFM segments, which allow you to send targeted campaigns to individuals, are a lot easier to create than models trying to predict who could potentially be a VIP.

Segmentation also paves the way for predictive analytics. Building and defining segments well would also allow you to identify which variables were important in predicting a customer’s behaviour. It leads you to the questions that future modelling needs you to answer.

If your focus has always been data collection, then segmentation allows you to explore your data and gain a better understanding of your customers and their behaviour. This also allows you to see where any gaps in data collection are. Complex analysis would not be able to give you this information.

Sometimes the simplest solutions are the best

Segmentation is not the fanciest of tools available in a marketer’s arsenal. But it is most definitely a very useful tool. If something simple and quick can triple your ROI, why not use it as your stepping stone to improving campaign performance? It is also important to remember that while predictive analytics allows us to peek into the future, we wouldn’t know what we were trying to predict if we didn’t have a good understanding of our database. This understanding is cultivated from the simple analysis that’s undertaken while building segments.