Customer analytics start with RFM

With more and more start-ups being built in the space where businesses sell products and services directly to consumers, who are also the end-users of its products or services, these ‘B2C’ companies now have much-valued customer data.

The way companies collect, aggregate and analyse this customer data will be a critical factor to determine whether these companies will survive and the trajectory of sales growth which they will be able to pursue. Customer data in today’s time is mostly collected online since most businesses sell online. For companies which sell offline, this is obtained via POS machines at the stores, indexed on the customer’s loyalty cards, and fed into centralised systems. Either ways this data is available for the company to draw key inferences, tweak offerings, decide customer engagement and plan marketing communications.

Irrespective of the size (revenues) and state (extent of technology deployment) of such businesses, every company can use this customer data to run some basic analytics. These analytics will help them manage their customer cohorts more effectively and also deploy their marketing budgets more efficiently.

One of the oldest, most basic and a very valued concept in customer analytics is RFM. The acronym RFM stands for three actionable parameters to analyse and segment customers:

• R for Recency of customer’s purchase
• F for Frequency, i.e. how often did the customer buy
• M for Monetary Value, i.e. how much did the customer spend

The interplay of these three parameters across customer cohorts helps understand the robustness of the customer base and design customer engagement interventions. These days customer acquisition costs are spiralling and hence instead, if equal attention is given to retention, it will be a bottom-line accretive decision both for the immediate and long term as a result of more loyal customers.

To run customer analytics, you could follow either of the two methods:

1. A simple analysis by arranging the customers in the descending order of each Recency, Frequency and Monetary value and then breaking these into manageable segments, say A, B and C, i.e. three segments. Then, you design independent strategies for each R, F and M to drive up each of these parameters. E.g. customers who have not bought from you for specific months (Recency), may need to be engaged to retain them or regain them. In contrast, customers in the C segment of Frequency, i.e. with lower buying frequency, will be targeted to ensure that their buying occasions increase.
2. The more sophisticated analysis is to arrange all the customers in the descending order of each parameter, say A, B and C segments for Recency, Frequency and Monetary value and then create a data cube with twenty-seven segments in all. You can then design strategies for each of these twenty-seven customer segments. This analysis is much more valuable and will certainly give a better ROI on marketing expenses. In the absence of sophisticated analytical tools, you could attempt to run the analysis in excel (provided data is not too huge) by organising the data, as shown in the figure below.

Questions to answer when you run RFM analysis:

• At the most basic level, are issues such as Who are your regular customers? Which of your customers are you most likely to lose? Who are your big spenders?
• How many one-time customers do you have?
• Which are the customers who have not bought most recently — could this information be used to predict customer retention rates?
• How many days after purchase (last buy) is the customer most vulnerable (likely to be lost)?
• How is the frequency of buying tracking over time?
• How is monetary value tracking? Is there an opportunity to upsell to the customer?
• How can you better predict the lifetime value of your customer, using RFM data?
• What kind of different campaigns/communications can you run based on which of the three parameters you want to target?
• Can your landing pages or checkout pages be designed differently for each customer segment, segregated based on R, F and M?
• Which of R, F or M, could help you improve your unit economics?
• How can you use predictive analysis based on this segment information to improve forecasting and reap the cascading benefits in your supply chain costs?
• Even if you are just a small business with a handful of customers, how can you be more effective in your conversations with your customers based on what RFM analysis tells you?

I am sure that the RFM analysis will give you some great insights and takeaways from your customer data. Attempt it to see!