How RFM analysis helps in customer segmentation?
RFM Analysis
RFM Analysis


RFM Analysis


If you think from a customer’s perspective then buying is not an easy process. One undergoes a series of long decisions taken by billions of neuron cells present in the brain and nervous system just to decide whether or not the product is worth the money to be spent. It’s strenuous to even think how that process works.


So, how can a seller know what his customer wants?  And upon knowing how can he be sure if the customer is going to buy the product or not? 


While there is no guarantee that a customer is going to buy a product, there are certainly a few things that can make selling easy for you. But for that you will first have to understand how an average customer buys.


How does a customer buy?


Curiosity is a big boon that mankind has been bestowed upon by the grace of good God. And potential curiosity leads to inquiry. The process of buying starts when the customer starts inquiring about your product. These inquiries usually expand over a spectrum of need and monetary value.


The next step is to compare, where a customer, after inquiring about a few similar products, decides which product to buy. He takes under consideration his buying range, advantages over other products and, depending upon the product (say a bicycle), life-span.


After the customer has compared enough products, he shortlises a product. He becomes your customer when he decides to purchase it.


Once purchased, the evaluation process comes to play. Here the customer uses the product and decides if it’s actually worth the hustle. At this point, it is decided if the customer is going to buy it again or not aka is he a loyal customer or not?


All the products are bought in this 4 step process, the only exception is luxury where the customer is tingled by delight. He buys products for his social standard and not need. So money doesn’t matter. Example: an Audi is not a need. You can travel in a basic car or bus or scooter too. But an Audi is a necessity when all your friends have BMWs and you want to look rich too.


All this seems to be quite simple, except it is not. It depends on the type of product or service you provide. You can’t expect a customer to buy a new mobile phone every month. Once bought, he is going to use it for at least a year. And if the product is a refrigerator then this usage will be 10-15 years.


So, how do we decide what to sell, when to sell and more importantly whom to sell?

Now, we have come to the point where we disclose the secret to this riddle.


RFM Analysis


R stands for Recency, F for Frequency, and M for Monetary value.


Let’s understand this.

  • Recency: When was the last time the customer visited your product section. It doesn’t matter if he made a purchase or not. Just visits count. It is highly relevant to show ads to a customer who was a recent customer in your product section.
  • Frequency: The total number of times the customer visited (within a specific time period that you measure from a certain point.)
  • Monetary: How much did the customer buy? It decides the purchasing power of the customer which is heavily affected by the kind of product bought.


Based on these three factors, we make out a ranking between: 1 to 5; where 5 being the highest score given to the best customers and 1 being the lowest given to the customers on whom you need to do a little extra work.


Depending on business and products we can either:

  1. Simply average the three factors and give a ranking.
  2. Or we can be smart enough to give weightage to the three factors and then calculate the appropriate rank.


Let’s discuss RFM analysis in detail. RFM analysis can be modified by giving value-based weightage to the three factors depending on what seems more relevant in your case.

Let’s consider 10 customers. Take recency for the number of days ago the customer visited. Frequency the number of times he visited or number of transactions (in the last 30 days) and Monetary the total money spent. 


Customer ID Recency Frequency Monetary
A1 13 5 540
B2 5 20 1029
C3 1 15 3200
D4 30 1 2390
E5 10 2 232
F6 21 10 453
G7 19 30 870
H8 15 4 2100
I9 2 20 2000
J10 12 13 500


Without wasting much time, let’s calculate the RFM score.


Recency score:


Rank and score on the basis of recency:

The customers who visited most recently are given preference. Hence C3 who visited yesterday has a score of 5 and D4 who visited a month ago has 1.


Customer ID Recency Rank Score
C3 1 1 5
I9 2 2 5
B2 5 3 4
E5 10 4 4
J10 12 5 3
A1 13 6 3
H8 15 7 2
G7 19 8 2
F6 21 9 1
D4 30 10 1


Frequency Score:


Ranking and score on the basis of Frequency:

Based on the number of times the customer visited your product section, he has been assigned  a score. G7 with 30 visits has the 1st rank and a score of 5. D4 on the other hand just visited once and thus has a score of 1. 


Customer ID Frequency Rank Score
G7 30 1 5
B2 20 2 5
I9 20 3 4
C3 15 4 4
J10 13 5 3
F6 10 6 3
A1 5 7 2
H8 4 8 2
E5 2 9 1
D4 1 10 1


Monetary score:


Ranking and Score based on Monetary value:

C3 spent a whopping 3200 bucks, the highest hence his score is 5 while E5 just bought 232 bucks worth of product and hence has a score of 1. 


Customer ID Monetary value Rank Score
C3 3200 1 5
D4 2390 2 5
H8 2100 3 4
I9 2000 4 4
B2 1029 5 3
G7 870 6 3
A1 540 7 2
J10 500 8 2
F6 453 9 1
E5 232 10 1


Based on these factors we will decide: Who is your best customer and who is the lowest?

But first assume that you are a cosmetic seller. You can give equal weightage to RFM factors and decide your loyal customers; pick your top 3 and bottom 3 customers. 


Giving equal weights mean that all the three factors are equally important for your products.

Down below is the table that shows RFM scores separately and then their average. 


Customer ID Recency Score Frequency Score Monetary Score Total Score: (Average)
A1 3 2 4 3
B2 4 5 3 4
C3 5 4 5 4.66
D4 1 1 5 2.33
E5 4 1 1 2
F6 1 3 1 1.6
G7 2 5 3 3.33
H8 2 2 4 2.66
I9 5 4 4 4.33
J10 3 3 4 3.33


Now, based on the score in the previous table, let’s give ranks to your customers.


Customer ID Rank Total Score: (Average) Customer Type (Considering all the scores, you may refer back to the previous tables)
C3 1 4.66 Loyal, can buy top quality, costlier products.
I9 2 4.33 Loyal, can buy top quality products.
B2 3 4 Loyal, but with low or medium budget products.
G7 4 3.33 Likes your products, interested in cheaper products.
J10 5 3.33 Can buy costly products, show him relevant ads.
A1 6 3 Can buy costly products, show him relevant ads.
H8 7 2.66 Can buy costly products, seems a little disinterested
D4 8 2.33 Doesn’t like wasting time, rich buyer; show accessories.
E5 9 2 One time buyer, not an interested customer
F6 10 1.6 Customer that is no longer loyal 


Please consider the following statements:

  1. Customers C3 and I9 like your products so they visit your product section quite frequently and have done a shopping of 3200 bucks.
  2. Customer D4, ranked at 8th on the other hand, came once and bought products worth a whopping sum of 2390. But he never showed up again. Show him ads related to accessories of the product he bought.
  3. Consider F6 ranked the lowest, bought only for 453 bucks but visited your store 10 times before doing so. He hasn’t been recent to your product section. This person is a careful buyer, with good quality products in the medium buying range he can become one of your best customers.
  4. G10 and J7, are your average buyers. You can lure them to loosen their pockets and buy more from you.


What does this tell you? This data boasts about your loyal customers who are more likely to buy your brand’s products. And at the same time, it also shows one time customer who does not seem to be that interested.

But this data is only for a month. All these numbers can change if the data is annual. Maybe annually D4 becomes your best customer as he buys big monetary value products every month without juggling much. 

It’s all about your marketing strategy and customer interests. This data seems more relevant now to decide your customers. 



What if your company sells mobile phones? They won’t shop again and again. As discussed above, you will have to weigh the recency and frequency more than the monetary factors.  

  • The customers who are more recent and frequent would be interested in buying from you. If they haven’t bought anything yet, you should run an advertisement for them as they seem to be the potential buyers. 
  • The customers who have bought from you and are quite recent on your mobile phone section might be interested in buying earphones from you.
  • The customers who are one time buyers are not that relevant to you. They might be interested in a different range of products from you, say cosmetics.


At accelmatic, we monitor this data precisely and based upon the type of products you sell, we decide the potential customers for you so that you don’t have to waste your money on someone who is not going to buy anything from you. In short, we will make the best of your budget .

The above clustering was a basic one depending on the RFM score. But RFM clustering isn’t that precise and more importantly can not predict future activity.

If we play smart then we can cluster customers better. This can be done by taking the following factors into consideration: age, gender, interests, country, spending habits, religion, race, income, education, lifestyle, etc. 


With this analysis, we can help you with:

  • Better customer service
  • Better targeted ads
  • Best marketing strategy


A teenager in New York, the USA and a senior citizen, somewhere in Delhi, India would have different interests but they can both be your customers if you sell them relevant products.


To conclude, the RFM analysis can help you segregate your customers. But it can work better if we consider practical factors that affect product sale. It can’t predict future actions but it can definitely figure out a potential buyer. With accelmatic, your smart marketing companion, you can monitor your data better and sell your product to the best of your market.