This post is our first in a series about recency, frequency, monetary (RFM) analysis, and why understanding how to leverage its concepts with data is crucial to your online business.

Suppose your commerce business has 20,000 all time customers, contributing $4,000,000 in gross merchandise value (GMV) over a three year period. You sell luxury jewelry and accessories – mostly online, but more recently in retail locations too. This year, you’re ready to embark on your first formal direct marketing effort. You already have a budget, and have hired a marketing manager to develop the plan. The key question becomes, which customers do we target to maximize the ROI of the marketing effort?

Intuitively, there are many decent answers to this, like targeting the customers who…

• Purchased your most expensive product once
• Purchased more than one product
• Purchased any product, but in the last 30 days
• Have a lifetime value in the 90th percentile of all customers in your database
• Have crossed over to a different channel, like from online purchasing to brick-and-mortar purchasing
• …

It’s easy to come up with a page-long list like this. But for your business, some of these options will prove more effective than others. So, what’s the best approach to understanding which customers might be most responsive to spending your time and budget to acquire? Of course, the answer lies with data (shocker!). How you interpret and act on your transactional data, though, is crucially important. Here, I would like to introduce a classic model for understanding where valuable customers come from – RFM analysis.

What’s RFM analysis?

RFM stands for recency, frequency, monetary analysis. It has its origin in the heavily brick-and-mortar retail industry, but has been adapted very successfully to online businesses over the last decade. At the foundation of the model is the assumption that 3 customer traits identify a customer as being valuable:

• Recency: has the customer made a purchase recently, relative to other customers?
• Frequency: does the customer have a predictable cadence of repeat purchases, relative to other customers?
• Monetary: does the customer have a higher average order value, relative to other customers?

If you’re not familiar with RFM analysis, its Wikipedia entry is a great primer. In addition, its effectiveness as a model for informing your business’s marketing strategy has been examined extensively as well. There are various ways to create an RFM “score” or “weight” to rank valuable customers over one another.

Understanding recency

Let’s go back to our jewelry business example. Of our 20,000 all-time customers, some of these made their last purchase yesterday. Those customers are impressionable, engaged with your brand, and it’s more likely that they’ll be retargeted by an ad campaign that you’re running. All of these points are consequences of recency — and it makes the customer more likely to buy again in the near future. There is also a subset of customers, whose most recent purchase was two years ago or more. These folks are no longer engaged with your brand, and for whatever reason, they have churned from your business.

Using data in your transactional database helps you put numbers to these patterns. The answers to each of these questions can be found easily in your DB, and you should learn them if you don’t know already:

• What percentage of your all-time customer base has made a purchase in the last 60 days?
• What percentage of all purchases from the last 60 days are from first-time buyers, vs. returning buyers?
• What are the most popular products purchased in the last 60 days?

Options to take action

It’s easy to understand that there’s a spectrum of customers based on recency: the more active recent buyers, and the inactive old buyers. For a marketing manager, this represents an interesting choice when choosing who to target with a limited budget:

• Should I target the more recent customers, since marketing to them makes keeps them engaged, and increases the likelihood that they return?
• Should I target the old inactive customers, since there may be a larger number of them, and since marketing to them may entice some to reactivate and become repeat customers again?

This choice is an opportunity cost problem. And it’s tricky. It might be the case that some of these old inactive customers never return if you target them, so you shouldn’t waste any marketing dollars on them. It also might be that the case that the recent, engaged buyers will continue being engaged regardless of whether they’re targeted – that’s another way to waste money.

So, how can we predict where our dollars are most effectively spent? Learning more about the frequency and monetary pieces of RFM analysis drives this action more.

Keep a look out for the sequels to this post for more info. If you’re currently an RJMetrics user, please reach out to our support team to see how your analyze your existing data in this way. If you’re not a user, but your business is searching for answers like these with a world-class product and team – please signup for a demo to find out more.