RFM Segmentation

In continuation of a series of posts on segmentation, let's talk today about RFM analysis, which is also based on the behavioral factors of a group or a segment of existing customers. The RFM analysis method allows you to assess general state of database, organize direct marketing competently and differ from competitors favorably. With correct motivation of the customer, it will be possible to improve the results obtained in the analysis. If you already know what the RFM analysis is, you will be interested in reading about our experience and discoveries related to the topic of RFM analysis.

RFM Segmentation Consists of Three Parameters:

Recency (R) — recency of last purchase

That is, how much time has passed since the interaction with the customer in days, weeks or months. Calculated as the difference between current date and date of last order. Customers, who recently made purchases from you, are more prone to re-orders than those, who have not shown any action for a long time. Users, who purchased long time ago, can only be renewed with offers that attract them to come back.

Frequency (F) — total purchase frequency

Shows how many interactions (purchases) for a certain period of time you had with the customer. If both sides are satisfied - there is a chance to maintain the frequency of purchases or increase them in your favor. The more customer made purchases from you, the more he is likely to repeat them in the future. Usually, this indicator is closely related to recency of purchase.

Monetary (M) — volume of purchase

Like previous indicators, is calculated for a certain period or number of interactions. Shows what was the “cost of customers” in terms of income and profitability, or rather, the amount of money that was spent. Grouped by monetary indicators analyses often get an idea of customers whose purchases reflect a higher value for your business. All the above indicators are important to calculate over the period that most accurately displays the desired data. Suppose you can take a sample for one year and divide it into quarters. Typically, a small percentage of customers respond to general promotional offers. RFM is an excellent segmentation method for forecasting customer reactions and improving interaction, as well as increasing profits. RFM uses customer behavior to determine how to work with each group of customers. The importance of indicators is ranked according to the sequence of letters – recency, frequency, monetary. Sometimes the name RF segmentation occurs when Monetary indicator is not used, because its indicator often depends on Frequency. Segmentation of customer base on this principle allows you to identify those with whom you really need to work, dividing them into customer segments (active, sleeping, growing), developing targeted marketing offers for the most active group of customers.

Segment your contact base easily

Drive Results

Definition of RFM Conditions

Definitions “recency”, “frequency” and “monetary” are understandable even intuitively, but our task is to turn them into figures that can be used to evaluate RFM, and this is somewhat more complicated. We will present our data in a form of a table sorted by the first value - date of last purchase (R).

Contact

Date of last purchase (R)

Volume of purchase (F)

Amount of purchase, UAH

Average bill, UAH (M)

Ivan Petrov

05.09.2013

5

2600

520

Pyotr Ivanov

07.08.2013

1

960

960

Oleg Pliushch

15.07.2013

3

1500

500

Sidor Petrov

04.06.2013

2

600

300

Olya Sidorova

25.02.2013

3

900

300

Anna Volkova

05.10.2012

4

3300

825

Immediately put on each indicator (R, F, M) “weight” for each customer based on the data obtained. To evaluate the customer base, we will use numeric values from 1 to 3, or in the percentage assigned to each customer as a result of the analysis. For convenience, we will divide the entire customer base for the beginning into 5 equal parts for each of the indicators. Suppose, from our example on an indicator “volume of purchase” F - 1,2,3,4,5.

1 - the worst for us volume, we will mark, as 1;

2,3,4 - the average result, we will mark it as 2;

5 - is the best value of F. This is our 3;

So: 1 is bad, 2 is average and 3 is good. Putting on each indicator “weight”, to use these weights to rank the list.

Contact

Date of last purchase (R)

Valume of purchase (F)

Average bill, UAH, грн. (M)

Weight for (R)

Weight for (F)

Weight for (M)

Ivan Petrov

05.09.2013

5

520

3

3

2

Pyotr Ivanov

07.08.2013

1

960

3

1

3

Oleg Pliushch

15.07.2013

3

500

2

2

2

Sidor Petrov

04.06.2013

2

300

2

1

1

Olya Sidorova

25.02.2013

3

300

1

2

1

Anna Volkova

05.10.2012

4

825

1

3

3

Now we can easily determine that for us the best customer with the results of 333. 111 shows that the customer is interested in us rarely, maybe even once. Based on the results you can choose how to deal with a particular group of customers. It's sad that usually 111 is the largest segment. And it's great that you cannot waste time on those, who are already lost, and concentrate on customers, who are really important for us. Another advantage of RFM segmentation is also in the fact that analysis can be done even for one indicator, which you are most interested in or combine indicators, although a full segmentation of the customer base will give you much more opportunities. Suppose we take as a basis only recency and frequency, and depict the resulting data graphically:

Green sector 5% - the best customers, who actively react to everything, buy, etc., respectively, sector 1.1 – ‘we are losing them”. You should work with each of the segments of the table differently, offering them different terms of cooperation. We always say that it is good to see the situation in static (how we do now), but it is more important to see in dynamics (where do we go). If you calculate the same table for the previous period and “overlap” it with the current one - you can see how the data is changing:

In sector 1.1, the indicator fell on 6%, due to decrease in number of passive customers. But in sector of 3.3, the number of “good customers” increased by 2%. Well, then we are working in the right direction. It is necessary to analyze due to what it happens and to fix the result. These data are already enough to work effectively with customers, but if you add a monetary segment to this indicator, then working on numbers will become even more interesting:) RFM allows you to segment the database so that you spend time and money on right customers. Try to do segmentation at least by one indicator, and even working with this data can promote the growth of regular customers. Next, we still have eRFM and a cohort analysis. Follow us on our pages on facebook or vKontakte, and also subscribe to the newsletter to be the first to find out all the fun;)

Want to conduct RFM analysis of contact base of your business customers? 

Get professional expertise

🔒 GDPR, CCPA, CASL Compliant. Your data is safe and secure with us.