Use RFM Customer Segmentation Model to Increase Your Sale
RFM customer segmentation model has been around for many years but it is not seen often in digital activities. And what a shame!
Author: Jacob Kildebogaard
There is tremendous knowledge in taking a closer look at your customers. RFM customer segmentation model has been around for many years and a widely used segmentation model in the mail order era, but it is not a model and approach that is seen so often in digital activities. And what a shame! I want to change that, because I know what an eye opener it is every time we do this for a client, and you should use RFM as well.
RFM stands for Recency, Frequency and Monetary, ie. customers are segmented based on:
- When did the customer last purchase (Recency)
- How many times have they purchased from us (Frequency)
- How much money did they spend (Monetary)
Customers are divided into 4 levels for each of the 3 variables, with 1 as best score, and 4 as lowest. That means a customer score on 111 is a customer who recently purchased, who have purchased many times and spent a lot of money, while a 444 profile is a customer who bought a long time ago, only bought once and spent a low amount. The segmentation is thus relatively simple, but provides great insights.
The value of RFM segmentation
So what do you get from segmenting your customers? The findings our customers typically find most valuable are:
- Knowledge of the distribution of customers in the various segments
- Data on permission delivery in relation to customer group
- The value of improving performance in digital marketing as RFM segments are actively used as audiences.
- 1 to 1 communication on email based on RFM location and segment change.
1. Customer Distributions
Do you have many loyal customers? Or do you actually have scary few who are really good customers? It is extremely important to know and act accordingly. The customer distribution are exciting insights that tell a lot about the business and what strategy you should use.
In addition to the knowledge and insight of your customers and business, you can also use it actively. You can treat them differently when buying. Your best clients should have VIP treatment and get a personal greeting about how cool they are to shop at you again. Though with a huge amount of buyers? It is all about automizing this part also.
2. Profile and permission share on the different segments
When you break down the segments into different variables you can see if there is a difference in the profile of the different segments – are your really good customers, for example, more wealthy? Or are you younger? It is valuable knowledge to use in your marketing so you can get more of the most valuable customers.
You can also see the percentage of customers with a marketing permission. It has been surprising how much it differentiates in the share with permission in relation to the RFM segment, and in some periods, it may also happen that first-time buyers very rarely gives permission. You can help this out by running a lead ad campaign on Facebook to users who have seen your receipt page (remarketing audience).
3. Active use of RFM segments in marketing
In addition to knowing the profile of the best customers and using it actively, you can also incorporate data customer data directly into your marketing. A smooth setup makes a dynamic update of the RFM location, as well as a change of the RFM segment, and send audiences to Google, Facebook and your email automation platform. This means that you can run different campaigns to the different segments, with the right messages. First-time buyers and regular loyal customers should not have the same creative, which would have been the case without this segmenting.
In addition to targeting specific customers, this data and audiences can also be used for look-a-likes and simular audiences. The most valuable segments are the ones you would most like to find similar customers, so full thrill on finding new customers that are similar to them.
4. 1-1 communication by mail
RFM segmented audiences for Google and Facebook are a great opportunity that brings results. But they are limited by the whole group getting the same messages, whether the user has just joined the segment or been there for eg 20 days. Therefore, communication via mail is perfect for 1-1 communication that can target precisely with the user’s data. One concrete example is when a user switches from having bought recently (R = 1) and they move down to R = 2, ie it has now been a little longer since they bought. The change should trigger an email, with the aim of reactivating the customer and making them buy. We know exactly how long since they last bought, we can integrate data on what they bought, and at the same time we know how much they typically buy for – all the knowledge that can be used to create the best possible content for success with making them buy again.
What preprocessing requires an RFM segmentation?
Before deciding to divide your customers into 4 segments for each of the RFM variables, consider how the breakdown should be. When did you buy recently? How many purchases are many purchases in your business? A standard does not exist as it depends a lot on your business – for some, it is very important that the user returns every year (eg Christmas) for others the goal is to increase the repurchase rate from quarterly to monthly (eg clothing).
Another consideration and preparation is how many years of data you should include. Does it make sense to try to reactivate users who bought 3-4 years ago, or should you rather focus on the last 2 years of data?
We also find that outliers can greatly affect data, so be sure to ensure data quality and investigate outliers and sort them out if necessary. An outlier is data located on the outer edge of the data set, for example, users who have bought profusely many times in your store.
When choosing a subdivision, the best process is to make a good proposal based on what makes sense for your business. Once you’ve divided your customers, you can evaluate the segments if they make sense or whether you need to make some definitions.
- The Monetary variable can advantageously be data for profit rather than revenue as it provides input on offer riders rather than product size. However, it may be more complex data to get into play.
- If you do not have a high volume of transactions, you may need to group the segments more, to meet audience size requirements at Google and Facebook. However, you will always be able to make a 1-1 mail
- The setup runs marketing that requires permissions. It is therefore important that you have linked data directly from unsubscribe to your setup in order to be up to date and legal.
So how do you get started?
If you want to get started yourself, start making the analysis manually, and find out how to define the different groupings. When you have this in place, and you hopefully see the advantage, it is about automizing it and making a smooth setup sending audiences to Google, Facebook and you mail platform automatically.
RFM segmenting is a great way to use your data actively to increase conversions and sales. Even if you have a great rate of returning customers, this model helps you find the segments with largest potential, and make your communication more relevant.
This post is written by Jacob Kildebogaard, one of the leading digital marketeers in Denmark. Jacob is media director and partner in Ambition, and you can hear him talk more about RFM setup in BigQuery to inOrbit 2020.