Ecommerce Growth, Retention Rate Optimization

What Everybody Ought to Know About RFM Analysis in eCommerce

full disclaimer before we get started: RFM analysis stands for Recency, Frequency, Monetary analysis and it has something to do with your customers

I love coffee. When I was living in Italy, I developed a habit of having one macchiato every morning before work.

What I found interesting about coffee in Italy is the fact that you can choose from a variety: espresso, cappuccino, marocchino, americano, stretto, ristretto, cafe latte, macchiato. It depends on your preference. We are all different and enjoy coffee in our own special way. Italians know this and found the secret formula for keeping us happy: tailor a product based on customer’s preferences.

What has coffee to do with customer engagement?

One day, as I was walking in the bar next to our office to ask for my regular macchiato I heard one of the baristas say:

“Ciao Bella, il tuo machiatto e pronto” / “Hi gorgeous, your machiatto is waiting for you”.

I had just entered the bar and my machiatto was already there on the counter, ready and waiting for me. Big smile on my face!

The baristas knew me. They knew approx the time of day I would walk in and what I like to drink and proactively made me an offer I couldn’t refuse. They transformed a normal, repetitive experience into an extraordinary and personalized one by saying “I know what you need. Here it is, waiting for you”.

They made my day. And they did something more than this. In marketing terms, what the barista did through his action was to : (1) generate positive emotions, (2) form a habit and (3) upscaled me as a customer.  After that experience, I kept coming back to them for my morning coffee. Besides my usual machiatto, I would add a “brioche con crema” (sometimes on the house) and also started taking my lunch and dinners there (late office hours).

I became a full customer for them.

Coming back to Earth, my barista story can’t scale.

Doing 1:1 marketing is not a feasible option. You may not be able to know each customer up close and personal (plus that would be a little creepy).

But that doesn’t mean you can’t start to know them through their buying behaviors. This can help you begin a relevant conversation which may lead to a happy customer and a long-lasting relationship.

First things first. In order for you to have a relevant conversation with your customers, you first need to know them.

How often have you, as a Marketing Manager, addressed the following questions:

  • Who are the customers spending the most?
  • Who are the most loyal customers coming back and placing a second, third, fourth order?
  • Who are the newest customers?
  • Who are those customers I am about to lose?
  • Who are those customers I already lost?

RFM analysis is a pretty cool way of answering these questions. And it’s not just me saying that. RFM analysis goes hand in hand with customer centricity, segmentation, personalization, tailored marketing campaigns and customer lifetime value. And the internet is hyped on all these indicators. It’s on everyone’s lips:

Know and understand your customers will be this blog’s mantra.

RFM analysis

RFM analysis (Recency, Frequency, Monetary) is…

…a marketing method that can help you, as a Marketer, segment your customers based on their purchase history. By answering these 3 questions: (1) how recently, (2) how often and (3) how much did a customer spent with you, you will be able to uncover buying behaviors.


Customers who recently placed an order on your website are more willing to place another order compared to those who haven’t bought in a while.

Customers that frequently buy are more willing to buy again rather than those who only bought once;

Customers that usually spend lots of money are more willing to buy again and generate revenues;

RFM is your ally in the retention game. When you know your customer’s buying behaviors, you are one step closer towards decoding their needs and preferences. So it is a step closer to starting a meaningful conversation.

According to an HBR article, retaining your existing customers is more effective for your business rather than going out and chasing after new customers.

“acquiring a new customer is anywhere from five to 25 times more expensive than retaining an existing one. It makes sense: you don’t have to spend time and resources going out and finding a new client — you just have to keep the one you have happy.”

Plus, it’s good for your company’s revenue (HBR article):

A 5% increase in customer retention can increase company revenue by 25-95%

Sounds great! As a marketer, where do I begin?

The first step is to have a look at your customer’s database and extract historical data available for calculating the recency – frequency – monetary values:

  • Recency – the date of the latest purchase made by a customer. What date is today? Deduct from this date the date of the last purchase made by a customer. You will get the number of days since the last purchase aka the recency. Smaller values indicate an active customer while bigger ones a dormant or lost one.
  • Frequency – the number of orders each customer placed with you up until today. Bigger values indicate a loyal customer on your website.
  • Monetary – sum-up the value of all the orders placed by a customer. Bigger values indicate a big spender on your website.

The second step: now that you have the data for all 3 dimensions, you need to group it and give each group a score. As I was doing a bit of research online, the most preferred way of doing this is to split the values for each dimension into 5 equal parts and assign each one a score.

From what I saw online, this is one of the pain points of working with RFM analysis but there are alternatives and things you can get inspiration from.

  • is offering an interesting approach to segmenting your data;
  • The Database Marketing Institute offers a good description as well;
  • At Omniconvert, we are also in the process of designing a dashboard focusing on retention KPIs. The dashboard will also provide an automated way of grouping data for you, based on your minimum and maximum values for each dimension (stay tuned for more). This goes hand in hand with the web personalization capabilities that we already have @ Omniconvert. Once you defined your segments you can go ahead and think of engaging experiments you would like your audience to see when surfing online through your products.

RFM score

Essentially, in the end, you need to come up with a score from 1-5 given to each dimension:

  • Recency – the highest score of 5 goes out to the customers who bought most recently;
  • Frequency – the highest score of 5 goes out to the customers who placed the highest number of orders;
  • Monetary – the highest score of 5 goes out to the customers who spent the most in terms of value.

You will get something like this: 555. A customer with a 555 RFM score is your most valuable customer. Call it a Champion or a VIP, they recently bought from you, placed the highest number of orders and bought of high value.

Look at RFM as a labyrinth. If you change one value from the equation, you are suddenly dealing with a new type of customer, with a different buying behavior, different needs so different marketing tactics.

The third point. Let’s play a game:

What does a score of 125 (R=1, F=2, M=5) tell you?

What you have here is a group of customers who (R) bought from you a long time ago, (F) they place few orders but (M) those were valuable orders.

What do you need to do here? You can’t afford to lose these customers! Maybe write a hand-written letter? Or maybe drop them a courtesy call to ask about their experience with your products and services.

What about a 442 (R=4, F=4, M=2)?

What you have here is a group of customers who (R) bought recently from you, with an (F) high frequency (so many orders placed) but of (M) low value.

What do you need to do here? Maybe run a loyalty campaign? Send them some discounts or coupons? Maybe recommend other programs?

So what can you do with this information?

It’s time to look in your backyard and see whether the marketing activities you have in the pipeline to execute in the next quarters are in sync with the customer segments you have in your database.

  • Are you properly rewarding you VIPs?
  • Are you providing differentiated discounts?
  • Are you touching base with your most loyal customers through a courtesy call to find out about their experience with the products and services you are delivering?
  • Are you building a relationship with your customer through the onboarding process?
  • Are you talking to those customers you are about to lose?
  • Are you creating tailored emailing campaigns?

Here is an example which may help you to put things into perspective when you think about how to make use of the RFM analysis, the points, and the segmentation.

The way you think and segment your customers will vary. You can play around with the points and create as many and as diverse groups of customers as you see fit.

This could be the beginning of a wonderful relationship…

We know your day is split between meetings, conference calls, talking to suppliers, preparing the next marketing campaign and looking at numbers from various sources.

What if you put everything on hold for a day and run a test.

Ask someone from your team to give you access to the information which will help you extract the recency – frequency – monetary information.

Once you have that, filter descending on revenue and recency (you don’t need Python for that!).

Add a pivot table in excel and see how many customers you have in your 1% of revenue.1% of the best customers generates up to 30 times more revenue.

Can you send an email to them to reward them? What would you write? Would you pick up the phone to ask about their experience with you? Maybe start with 5 or 10 customers to get a feeling of the real world

Any small step matters, don’t give up!

It may be better than blasting everyone with the same message which doesn’t bring them any value.

Like the other day, I received an email saying “Hi Ethan”

  • #1 – My name is Anca;
  • #2 – Ethan is a boy’s name and I am a girl;
  • #3 – Don’t come back with a reminder saying “Hi Ethan” when I already wrote to you giving my real name and pointing out I am a girl.

Coming back to my simplistic coffee example:

In the name of RFM analysis, I would be considered a “loyal customer” based on:

  • Recency – days since last purchase – I would get the highest score of 5 because, for my entire stay in Italy, I would regularly come in and get a coffee from that bar;
  • Frequency – the number of purchases made – I would get the highest score of 5 because I was having one and sometimes two coffees a day;
  • Monetary – the value of the purchases made – I would get a score of 2 because my purchases were not very valuable (money wise) compared to what I could have spent there.

Drawing the line, my RFM score would be 552, a score which makes me a loyal customer.

What did the baristas do with this information?

Seeing that I am a loyal customer and knowing what my favorite type of coffee is, they provided a personalized experience, proactively coming to my rescue.

They also recommended a “brioche con crema” to go with the macchiato (on the house).

They started building a relationship with me as a loyal customer by allowing me to pay later or the next day.

What were the results of this tailored approach?

A happy and loyal customer. Not only did I become a loyal customer for breakfast, I would also come back for lunch and dinner. I would go for a second cup and even bring in my colleagues.

As we created a connection, they soon found out my name and started greeting me with “Buongiorno Anca!” which created a sense of familiarity and community (much needed for a foreigner like myself).

I may be a sucker for personalized interactions but think about it: who wouldn’t like to be greeted with “Ciao Bella!”…or “Ciao Bello!”. I bet it works great both ways.

In closing, I am curious to know:

  • Have you heard of RFM analysis before?
  • Are you already performing an RFM analysis on your eCommerce customers?
  • Were there any findings?

Looking forward to hearing from you in the comments section below!

Leave a Reply

Your email address will not be published. Required fields are marked *