Best Practice Customer Segmentation Implementation for Brand Communications
Key marketing and business objectives can be achieved by implementing best practice customer segmentation strategy.
One of the things we see all the time is organisations segmenting customers in a way that is not only useless to an organisation, but also does not improve the relevancy of their customer contact.
This is partly because the organisation was not clean on why they were doing segmentation in the first place. It’s also because there was no obvious plan to execute once the segmentation was in place.
The first thing we always reiterate to our clients is that they should never forget the reasons why they should segment customers. It enables two key objectives:
- Communicate in a more relevant and personalised manner to their customers; in order to
- achieve business outcomes
Before doing any analytics, an organisation needs to agree that not all customers are created equal, they deserve to be treated in specific, pre-defined ways, and these ways are agreed to in advance. It’s all very well to identify high value customers on your database but useless if you’re not going to do anything special with them.
Segmentation must support action.
Once there is a good understanding of how the organisation wants to personalise their customer contact, there are plenty of ways that analytics can help. Using behavioural, transactional and 3rd party data allows an organisation to build current and predictive models for the customer base.
It needs to be personalised.
When thinking about personalisation, this can do done in four main ways.
Personalisation is about WHO should be contacted, WHEN they should be contacted, HOW they should be contacted (the method of contact) and WHAT should they be contacted with (the substance of that contact). Analytics can help define the rules of personalisation for each customer.
To illustrate, figuring out WHO should be contacted can be done by considering these features:
- likelihood to respond to a campaign (whatever contact the campaign contains)
- likelihood to churn
- likelihood to be a high lifetime value customer
- likelihood to be approved (if organisation is a bank)
Analytics can be used to predict the likelihood for all these things occurring for each customer. As long as there’s enough historical data (and a reasonable set of potential factors that can be used in the prediction), a model can be created with high prediction accuracy.
As mentioned above, the analytical work is only going to be useful if the organisation has agreed that they are going to set up the processes to act when the models says they should. There is no point predicting that a customer is highly likely to churn in one particular month when the organisation can’t act on this information for another 3 months.
Personalisation leads to learning.
A final point to make is that the organisation can learn a tremendous amount about their customers when setting out to personalise their customer contact. Analytics is a process of uncovering information about customers, common patterns and anomalous behaviour.
For example in the process of creating a model that will be used to identify those likely to churn, the business will define what a churner is, learn about the quantity of churners across the database in the past, present and future, and they will learn the factors that drive churn potential leading to other churn reduction initiatives across the business beyond customer communication.