RFM (recency, frequency, monetary) analysis is a marketing technique used to determine quantitatively which customers are the best ones by examining how recently a customer has purchased (recency), how often they purchase (frequency), and how much the customer spends (monetary). RFM analysis is based on the marketing axiom that “80% of your business comes from 20% of your customers. ” It is also known as “Pareto’s rule”. It means that, theoretically, if you can keep just 20% of your best customers, you would still earn 80% of your revenue. Why? Because these customers are mostly loyal and keep coming back.

Using RFM, you can segment customers into groups by following criteria:

Recency – days passed since a customer’s last purchase (the less the better)

Frequency – number of purchases the customer has made (the more the better)

Monetary – total money spent (the more the better)

You can imagine simplest RFM model as a Rubik’s cube. Let’s say you have 300 customers who have made at least one purchase. On each axis you will put one of the criteria – R, F and M. You have 300 imaginary customers in the cube.

On the R edge, there will be 100 customers who made a recent purchase, 100 customers for whom a moderate time has passed since their last purchase, and 100 customers whose last purchase was a long time ago. You will do the same for the F and M edges of the cube. This way you will get your customers distributed over 3 x 3 x 3 = 27 small cubes.

In the top rear of cube you will find your best, most loyal customers:

those with lowest Recency, highest Frequency and highest Monetary value. The diagonally opposite corner, at the bottom of the cube will contain your worst customers, who are probably already gone.

Let’s call the small cubes “segments” and try to imagine what would be the best approach for each of them. Writing a personal letter to the best segment with few people in it sounds pretty feasible and effective, right? You may want to encourage first-time buyers who have made a purchase recently to make another purchase. Or there might be some customers will buy from you again who may have just forgotten about your shop. A little reminder wouldn’t hurt.

Why it works so well

When somebody decides to make a purchase, he creates a stronger pattern that is more likely to repeat over your entire customer base. This model is surprisingly simple, but it works similarly well compared to complex neural networks or other advanced techniques based on many attributes. Studies have been made on what the influence that Recency, Frequency and Monetary value have on the likelihood of a repurchase. The strongest indicator wasn’t Revenue (or the Monetary value). This actually had the smallest influence, by far. The major indicator was Recency, followed by Frequency.

How “L” makes the RFM model actionable

“L” stands for purchase (L)atency. Purchase Latency is the average number of days between subsequent purchase orders. You’ll need to analyze your transactions to calculate that number. Add this number of days to the date of customer’s last purchase, and you will get an estimated date of his/her subsequent purchase. Why would you send somebody a general Thank You email right after his initial purchase if you know that the guy will most likely repurchase in (for example) 37 days? Wait 30 days after he places an order, and then send him the offer. Strengthen his need to act quickly by adding a discount coupon that’s valid for 7 days. 30 + 7 is 37 days, which exactly matches his probable repurchase time, according to his average purchase latency. If he doesn’t convert (i.e., make another purchase within 7 days), email him on day 37 with an even better offer valid that’s until day 44. Automate that process to trigger emails for relevant customers every day. You’ll soon see a spike in your revenue. Open and click rates will grow along the way, of course. The offers you send shouldn’t be predictable. If you teach people that they will receive a discount every month, nothing pushes them to act soon to place an order. If you start triggering the emails according to how people shop, they will get what they deserve. People quickly learn that. They will feel that they are being treated more exclusively, thanks to the limited number of people receiving that particular incentive at a given time. Especially loyal, growing and new customers are sensitive to too-early offers. They feel spammed and tend to ignore the offers. It’s better to wait for a while and then hit them with an attractive, hard-to-turn-down offer.

RFM(L) as Customer Retention tool

Business owners generally agree that it is about 5 times more expensive to acquire a new customer than to retain an existing one. That’s why you should treat your existing customers like gems.

However, hardly any business today has a clear strategy for customer retention. You probably understand how to send the same newsletter to all your contacts. I respect the fact that this is probably the best you could manage to do until now, but now I’m going to teach you much more efficient approach.

By that, I mean teaching you how to work with customer segments so that your targeted emails will be different from your common newsletter, so that you can maximize the response or conversion rate. It matters how you phrase and place the subject, salutation, email content, recommended products as well as promo discount type or other incentives.

We developed a set of automation rules and marketing recommendations for typical RFM segment groups. We recognize seven segments based on the RFM model.

A nice thing about Insightee segments is that they are disjunctive, which means they don’t overlap. By definition, it cannot happen that one customer would appear in more than one segment at the same time. The segments cover all customers. By “customer” I mean anybody who made at least one purchase during the analyzed time.

For the simplest, yet still-meaningful segmentation, we need just the (R)ecency and (F)requency displayed as a matrix. Let’s put Recency on the vertical axis and Frequency on the horizontal axis. Particular combinations of Recency and Frequency create a visual “grid”. When you calculate these parameters for every customer, it is clear which segment he belongs in.

In addition to Recency, Insightee calculates average purchase Latency (as explained above) and splits the Recency axis by the Latency multiples. That lets the algorithm accommodate to any dataset and still respect the natural “purchase cycle”, i.e. average Latency of any business. Simply put, this way you can calculate your segments more accurately.

The (M)onetary metric is then used to calculate an appropriate one-time discount that can be offered to each segment. Ideally you should know your average margin. Most businesses I spoke with have margins around 25%.

Let’s say, for example, that your customers spend an average of $49 per purchase (Average Order Value) and that first-time-buyers repurchase, on average, 23 days after their initial order. If your average margin is 25%, you can automate the sending of an email offer on day 20 after the first order, with a free-shipping discount on any order exceeding $49. You can state clearly that the offer is valid only for several days.

Such rules can then be automated for the other segments. Don’t forget to measure everything. Tools like MailChimp and Google Analytics will be your best friends for that.