Mike first cited Donald Libey’s famous book on RFM, published in the early ‘90s.

In short, customers can be defined by three metrics:

  • Recency: The freshness of the relationship between your brand and your customer; indicates when customers slip from active to inactive; the primary measure of business vitality.
  • Frequency: The measure of demand; measured in number of orders per period of time.
  • Monetary value: A measure of customer worth; measured as average order value.

RFM matrix

Here’s the RFM matrix that Mike used to find his hypothetical best customers. You’ll see he has included some numbers for each metric, as an example.

Of course, the numbers one uses here to can have a big effect on RFM analysis. Optimising how customers are categorised, finding the right values for R, F and M can be achieved using CHAID.

rfm matrix

Who are your heroes?

Once the matrix and the categories are finalised, the next step is to identify which customer sets are most important and how your efforts should be divided amongst them.

Here you’ll notice that Mike has indicated the hierarchy of value within recency, frequency, monetary value. Recency is the most important metric as it’s seen as most indicative of an active customer.

Of course, a recent and frequent customer is the ideal, whatever their spend. Lapsed heroes, who have spent big and regularly in the past but not for a while, may be worth spending some time on.

rfm matrix colour coded by customer value

Different heroes, different messages

Drilling down into customer data can help to establish some further metrics, aiming to firm up our appreciation of customer value.

Two customers that look pretty similar, when examining recency, frequency and monetary value, may in fact be two very different types of heroes requiring very different marketing messages.

Let’s take a look at an example that Mike gave:

two different hero customers

You can see here that the two customers have very similar values for recency (26 VS 33 days), frequency (6.5pa VS 8.3pa) and monetary value (£806 total VS £946).

However, key differences can be seen between the two customers, through the lens of gross margin (£187.5 VS £49.7). This looks to be explained by the propensity of one shopper to buy a high number of discount items.

This means that although both should be given love as important customers, the discount shopper should perhaps be targeted with bundled offers or incentivised to refer a friend. The shopper more inclined to buy at full price should perhaps be given more exposure to branding, new products and repeat purchases, or perhaps added value offers (e.g. giftwrap).

Ecommerce teams

Mike’s thoughts on data triangulation led to some conclusions, outside of the strategy for selecting best products, best customers and generating most profits.

The main one is that marketing and merchandising cannot exist in isolation in an ecommerce team. Knowledge of product and customer is essential for merchandising, media spend, email marketing and more. If these two groups of people aren’t exchanging ideas based on regular testing, success will be harder to come by.