22 Dec 2016

Gathering Customer Insights from Clustering in Tableau

Katelyn Weber

From websites to retail shops, businesses are using data to get a handle on who their customers are and how they behave. Gathering this data-driven insight enables services and marketing to be tailored to the customer like never before.

3 common sense KPIs for customer analytics

One cornerstone of customer analytics is finding sensible ways to group customers in order to profile them and understand the services they are using and generate marketing materials which appeal to them. Grocery outlets might want to know which customers actually use their discount coupons and whether offering these discounts yields a profit. Websites might want to know whether people who stay on the page longer also visit the site more frequently and how that feeds into the likelihood they will make a purchase.

A common approach to this type of customer analytics is called Recency-Frequency-Value (RFV) analysis. This approach is pretty simple: create three key performance indicators measuring

  • Recency – how recently did this customer last engage with us?
  • Frequency – how often does this customer engage with us? Are they a regular customer?
  • Value – how much does this customer spend with us? How valuable are they to us?

From KPIs to Insights with Clusters

The challenge comes when we want to combine these three KPIs into customer profiles to build insights and make decisions about how to target these customers. How do we decide which customers are the most similar to each other? Clustering provides a data-driven method of sectioning records (be they customers, products, patients, transactions, etc.) into similar groups. There are lots of clustering methods, but one easy way to get insights quickly is by using Tableau 10's new one-click clustering capability. By dragging and dropping "Cluster" into the view, Tableau will automatically group your data into clusters of similar customers which you can save to dig deeper and find each group's common characteristics. Finding averages of the Recency, Frequency, and Value of each cluster can provide you with a quick way to profile these customers within their groups.

Marketing - RFM Visual Drag and Drop in Tableau

Telling a story with your data

The real value in clustering comes from being able to use the average Recency, Frequency and Value metrics for each cluster to tell a story about who the customers are within each group. Think about how a "typical" person behaves within each group in order to turn clusters into customer profiles.

Maybe you find a group of customers who were high spenders but haven't been to your website in quite a while; how can you reach out to re-engage these customers in your products? Or, perhaps some customers have been placing lots of orders with your delivery service, but they only do it when a huge discount is available. It could be time to revisit the types of discount offers these customers are given in order to improve profitability.

From these examples, you can see that turning figures into a story gives business leaders straightforward insights they can use to make immediate decisions. Clustering pairs perfectly with customer analytics to form these types of profiles, and now that Tableau offers clustering out of the box, it's as easy as a drag and drop to build clusters of your customers.

Interested in seeing clustering in action? For a hands-on example, flip through the Tableau dashboards just below.

Katelyn Weber

About the author

Katelyn Weber is a Consultant at Concentra. She specialises in Tableau and Alteryx-based business reporting and regularly trains analysts to use these platforms.

Subscribe to our newsletter

Get in touch
Get the edge

Go to webform

Subscribe to our newsletter

Get in touch
Get the edge

Go to webform