How we built an automated RFM analysis

posted in: Uncategorized | 0

…and lived to tell the story

Pssst, hey you!

Yes you, drowned in an ocean of data. Running reports and preparing presentations for the next marketing meeting. Thinking of what the hell should be done in the next quarter to turn things around and engage your customers better.

You know that the only way to get to your customer’s heart is by knowing who they are. But somehow, the most important action always ends up at the back of the list with priority zero.

There’s just too much focus on reporting and too little on understanding and engaging the people behind the purchases.

So, once the new quarter begins and the marketing budgets are being released, you stick with the same old strategy and pretty much just pray it’s going to work this time as well.


What if you stopped for a moment, put that “client-centricity” buzzword to work and asked yourself:


In fewer words, what’s your relationship with your customers? Is it a long-lasting love relationship, or are you heading into a nasty breakup with half of the customers in your database?

What if I told you our automated RFM analysis can help you shed some light over the love stages in your client database?


Or, what about…

Those loyal customers, who showed you constant love (R = 4 – 5, F = 3 -5, M = 3 – 5);

The young lovers who want to know you better (R = 4-5, F = 1 – 3, M = 1 -2);

The unfaithful Don Juans, who are flirting with the next hot thing in eCommerce (R = 1, F = 1, M = 5);

The ones who are about to dump you because you’re not paying attention to them (R = 2 – 3, F = 1 – 5, M = 1 – 5);

Those ex-lovers who already left you behind for the competition (R = 1, F = 1-5, M = 3-5);

And the bad break-ups who might want their money back (R = 1, F = 1 – 5, M = 1 – 2).

Do you know any of these? They are in your database, waiting for you to discover them, engage them and speak their language.

So how can we help?

No hero ever accomplished any mission without a trusted sidekick. Harry Potter had Ron Weasley. Here, at Omniconvert, we have the RFM analysis as our trusted, loyal partner in crime (a bit of a stretch here, but trust me on this and continue reading).

It was not too long ago that we realized the magical powers of the RFM analysis. We even spoke about it here and here (Valentin Radu is doing an awesome job explaining it).

In a nutshell, by knowing (1) how recent a customer bought from you, (2) how many orders they placed, (3) and the total value of their order, you can detect the love level your customers have for you. You can also create customer segments of different values for your business and with different buying patterns.

So we set ourselves on a journey to using this RFM analysis with our clients. And while on that journey, we first stopped and asked ourselves: how can we design a segmentation that saves time, is easy to use and gives instant access to the customer segments hidden in the database? Our objective was to handle the heavy lifting of number crunching and give marketers time to breathe and focus on creating relevant marketing strategies for each customer segment.

Through multiple iterations and brainstorming session, we designed an RFM analysis that is fully automated and built on each shop’s history.

Here is how it works

Our RFM analysis is part of a bigger product created for the Magento eCommerce platform, Growth Engine for Magento. So, if your customer database is in Magento, here is how the model works.

To start from the end, once our product has been installed in a Magento shop, the RFM model automatically generates customer segments. For each customer segment, we display the number of customers from each and their lifetime revenue so that an eCommerce manager can instantly know who to approach first.

If, for example, you know you have a lot of “Passionate new customers” in your database (R= 5, F = 1-2, M = 4-5) bringing an important chunk of your revenue, you may want to understand what made them choose you and what you can do to make them stay for more than a one-night-stand.

What is the maths behind this?

⇒ We start by detecting the data from your Magento shop. More specifically, we look at the minimum and maximum values for Recency (R), Frequency (F), and Monetary values (M) from your store;

⇒ We then split the data for R, F, and M into 5 buckets with the use of percentiles (see table below);

⇒ Each bucket will receive a score from 1 to 5 (where 1 is the lowest and 5 the highest) (see table below). For example, a customer who bought from you 1 year ago will probably get a low score for Recency, of 1 or 2 (depending on your database and your sales cycles). A customer who placed 10 orders may receive a score of 5 for a store where the sales cycles are longer and a score of 1 for a store where the sales cycles are shorter.

Points Recency (days since last purchase) Frequency / Monetary values
5 within the last month customers who are in the top 5%
4 within the last 3 months customers who are in the top 20%
3 within the last 6 months customers who are in the top 30%
2 in the last year customers who are in the top 60%
1 more than a year ago the customers who spent and bought the least 

 ⇒ Each one of your customers will receive points for Recency, Frequency, and Monetary based on buying pattern in relation with all the other customers;

⇒ After points are being assigned, each customer in your database will receive a unique score. This score will constantly change based on the customer’s interaction with your store. RFM score = 555 | RFM score = 234 | RFM score = 115 | RFM score = 313

⇒ We group similar scores into 11 RFM Groups and display them on the dashboard, where you can see them, analyze them, and make a further business decision based on your conclusions;

⇒ For each group and sub-group of customers we display how many customers there are in there and the revenue they brought so far.


We are not the boss of you. You are the boss of you. If you feel our customer segmentation does not do the trick, you can change it and adapt it in many ways possible:

⇒ You can change the way you assign points

⇒ You can change the way you create the customer groups

⇒ You can even change the name of the groups

What can you do with it?

Now you know who’s hiding in your database. Good, it’s time to act.

⇒ Once our product is installed in Magento and the customer groups are created, you can instantly filter from Magento a specific customer group you want to interact with and prepare a special emailing campaign just for them. 

You know you have a lot of “About to dump you” customers (R = 2 – 3, F = 1 – 5, M = 1 – 5). They are disengaging from you. Think of a re-engagement campaign that would bring them back on your website. Send them a personalized email and asked them what happened that they stopped visiting you.

You have your “Soulmates”. They are your ideal customers. Reach out to them and see your store through their eyes. Maybe think of re-designing your web experience with their help or reward them for being loyal to you for so long.

⇒ The RFM analysis from Magento is also integrated with the Omniconvert web personalization platform. This means that you can apply A/B tests, overlays/pop-ups or online surveys only on a selected customer segment. Our product has a native integration with Omniconvert platform allowing you to instantly jump from one platform to another and run your experiments.

Like for the “Fliter-ers” (R = 4, F = 1-2, M = 4-5), they are active placed 1-2 orders of high value. Use your charms and make them order more.

You have your “Apprentices” (R = 4-5, F = 1-3, M = 1-2), those new customers who are very active but new to your store so they are not spending too much. Prepare an online survey and find out more about what they are looking for.

As with any love relationship, your customers go through many love stages with your store. When you know which is which, you can give them their own special treatment in pricing, email campaigns and website experience based on the value they bring you.

Ok, If you are reading this it means you reached the end of the story.

Whether you scanned the article or actually read it, it doesn’t matter. If there is one thing you should take away with you is the need to reach out and know your customers for a long-lasting love relationship.

Are you ready to find the love hidden within the numbers of your customer database?

Let’s talk. 


I am leaving you with Valentin’s video, on what RFM analysis is and how it can help marketers like you.

P.S – Oh, yes must not forget. Stay GDPR compliant in everything you do! Play it safe!