RFM(E) analysis overview
Digital marketers start to understand the importance of “know your customer.” Instead of focusing on generating more clicks, marketers need to shift their focus to retention, loyalty and building customer relationships.
Instead of analyzing the entire customer base as a whole, it’s better to segment them into homogeneous groups, understand the traits of each group, and engage them with relevant campaigns rather than segmenting on their demographics like country, sex or age.
One of the most popular, easy-to-use, and effective segmentation methods to enable marketers to analyze customer behavior is RFM or RFE analysis. RFM segmentation categorizes your customers into different segments, according to their interactions with your website, which will allow you to subsequently approach these groups in the most effective way.
What is RFM Analysis?
RFM is a behavior-based segmentation approach, which stands for Recency, Frequency, and Monetary value, each corresponding to some key customer trait. These RFM metrics are important indicators of a customer’s behavior because frequency and monetary value affect a customer’s lifetime value, and recency affects retention, a measure of engagement.
Businesses that lack the monetary aspect, like viewership, readership, or surfing-oriented products, could use Engagement parameters instead of Monetary factors. This results in using RFE – a variation of RFM. Furthermore, this Engagement parameter could be defined as a composite value based on metrics such as bounce rate, visit duration, number of pages visited, time spent per page, etc.
RFM factors illustrate these facts:
- the more recent the purchase, the more responsive the customer is to promotions
- the more frequently the customer buys, the more engaged and satisfied they are
- Monetary (engagement) value differentiates heavy spenders from low-value purchasers or users
It helps managers to run an effective promotional campaign for personalized service and can helps marketers find answers to the following questions:
- Where to focus your efforts in order to receive bigger value?
- Segment your customers based on their behavior and patterns?
- Build personalized marketing retention campaigns?
- Who are your best users?
- Which of your customers could contribute to your churn rate?
- Who are your potential customers?
- Which of your customers can be retained?
- Which of your customers are most likely to respond to engagement campaigns?
RFM analysis details
Let’s demonstrate how RFM works by considering a sample dataset of customer transactions. Table 1 contains recency, frequency, and monetary values for 15 customers based on their transactions.
Table 1: Example Customer transactions dataset
RFM Analysis Example
There are many ways in which the RFM score can be calculated. Start with using the discretization method and then calculate the score using the arbitrary/expert split.
With discretization, customers are automatically divided into quartiles/quintiles (segments) of relatively equal size. The resulting four/five groups will all have almost identical numbers of customers.
- Arbitrary/expert split
With the arbitrary split, each marketer defines a fixed set of intervals where certain values equal certain scores. For example, R score 1 is granted to the customer who bought a product last time six months ago, R score 2 is granted to the customer who bought it three months ago. The challenge using arbitrary/expert split is that you need to choose the right values for the intervals and sometimes there are errors or wrong assumptions made.
We advise using values in the report on the discretization level. Assume that we rank these customers from 1-5 using RFM values.
Let’s begin with ranking customers on recency first, as shown in the below table:
As seen in the above table, we have sorted customers by recency, with the most recent purchasers at the top. Since customers are assigned scores from 1-5, the top 20% of customers (customer 12, 11, 1) receive a recency score of 5, the next 20% (the next 3 customers 15, 2, 7) a score of 4, and so on.
Similarly, we can then sort customers by frequency from most to least frequent, assigning the top 20% a frequency score of 5, etc. For the monetary factor, the top 20% of customers (big spenders) will be assigned a score of 5 and the lowest 20% a score of 1. These F and M scores are summarized below:
Finally, we can rank these customers by combining their individual R, F, and M rankings to arrive at an aggregated RFM score. This RFM score, displayed in the table below, is simply the average of the individual R, F, and M scores, obtained by giving equal weights to each RFM attribute.
Recency, Frequency, and Monetary Analysis
The next question that arises is: Is it fair to average out the individual R, F, and M scores for each customer and assign them to RFM segment, as per their purchase or engagement behavior?
Depending on the nature of your businesses, you might increase or decrease the relative importance of each RFM variable to arrive at the final score. For example:
- In a consumer durables business, the monetary value per transaction is normally high but frequency and recency is low. For example, you can’t expect a customer to purchase a refrigerator or air conditioner on a monthly basis. In this case, a marketer could give more weight to monetary and recency aspects rather than the frequency aspect.
- In a retail business selling fashion/cosmetics, a customer who searches and purchases products every month will have a higher recency and frequency score than monetary score. Accordingly, the RFM score could be calculated by giving more weight to R and F scores than M.
- For content apps like Hotstar or Netflix, a binge-watcher will have a longer session length than a mainstream consumer watching at regular intervals. For bingers, engagement and frequency could be given more importance than recency, and for mainstreamers, recency and frequency can be given higher weights than engagement to arrive at the RFE score.
This simple approach of scaling customers from 1-5 will result in, at the most, 125 different RFM scores (5x5x5), ranging from 111(lowest) to 555(highest). Each RFM cell will differ in size and vary from one another, in terms of the customer’s key habits, captured in the RFM score. Obviously, marketers can’t analyze all 125 segments individually if each RFM cell is considered a segment, and it’s difficult and overwhelming to visualize this imaginary 3D cube!
In general, the monetary aspect of RFM is viewed as an aggregation metric for summarizing transactions or aggregate visit length. Therefore, these 125 RFM segments can be reduced to 25 segments by using just R and F scores.
Create score-based segments
Categorizing the users and put them into segments is the last step of our effort. There are various approaches and we advise each business to review 125 segments and define its 11 categories, normally we conduct a workshop session in order to guide the process. The table below is something which we recommend to be used as a starting point and is not perfectly suited for all projects as your customer composition might drastically differ. Therefore, use the table as a starting point and if needed make the necessary adjustments.
|Champions||555, 554, 544, 545, 454, 455, 445|
|Loyal||543, 444, 435, 355, 354, 345, 344, 335|
|Potential Loyalist||553, 551, 552, 541, 542, 533, 532, 531, 452, 451, 442, 441, 431, 453, 433, 432, 423, 353, 352, 351, 342, 341, 333, 323|
|New Customers||512, 511, 422, 421 412, 411, 311|
|Promising||525, 524, 523, 522, 521, 515, 514, 513, 425,424, 413,414,415, 315, 314, 313|
|Need Attention||535, 534, 443, 434, 343, 334, 325, 324|
|About To Sleep||331, 321, 312, 221, 213, 231, 241, 251|
|At Risk||255, 254, 245, 244, 253, 252, 243, 242, 235, 234, 225, 224, 153, 152, 145, 143, 142, 135, 134, 133, 125, 124|
|Cannot Lose Them||155, 154, 144, 214,215,115, 114, 113|
|Hibernating customers||332, 322, 233, 232, 223, 222, 132, 123, 122, 212, 211|
|Lost customers||111, 112, 121, 131,141,151|
Based on the calculated RFM score, the customers are put into 11 segments:
|Champion||Bought recently, order often, and spend the most.||Reward them. Can be early adopters of new products. Will promote your brand. Most likely to send referrals.|
|Loyal||Orders regularly. Responsive to promotions.||Upsell higher value products. Ask for reviews.|
|Potential Loyalist||Recent customers, and spent a good amount.||Offer membership/loyalty program. Keep them engaged. Offer personalized recommendations.|
|New Customers||Bought most recently.||Provide onboarding support, give them early access, start building relationships.|
|Promising||Potential loyalist a few months ago. Spends frequently and a good amount. But the last purchase was several weeks ago.||Offer coupons. Bring them back to the platform and keep them engaged. Offer personalized recommendations.|
|Core||Standard customers with not too long-ago purchases.||Make limited-time offers.|
|Needs attention||Core customers whose last purchase happened more than one month ago.||Make limited-time offers. Offer personalized recommendations.|
|Can’t lose them but losing||Made the largest orders, and often. But haven’t returned for a long time.||Win them back via renewals or newer products, don’t lose them to competition. Talk to them if necessary. Spend time on the highest possible personalization.|
|At Risk||Similar to “Can’t lose them but losing” but with smaller monetary and frequency value.||Provide helpful resources on the site. Send personalized emails.|
|Losing but engaged||Made their last purchase a long time ago but in the last 4 weeks either visited the site or opened an email.||Make subject lines of emails very personalized. Revive their interest by a specific discount on a specific product.|
|Lost||Made last purchase a long time ago and didn’t engage at all in the last 4 weeks.||Revive interest with reach-out campaign. Ignore otherwise.|
Advantages and limitations of RFM-analysis
The chosen example shows that RFM-analysis can be used for a wide range of business units and business cases. The assumptions and attributes of the model can be changed in a targeted manner. For example, customers can be split up in advance based on certain characteristics (e.g., by industry) and the RFM-analysis can thus be performed on an industry-specific basis.
RFM-analysis can also be extended by the Length parameter (LRFM-analysis), which describes the number of days since the first contract was signed and how long the customer has been stored in the database.
Furthermore, the computed parameters are well suited as features for clustering methods from the Machine Learning area and thus offer a foundation for forecasting models. However, this shows the limitations of RFM-analysis, as it is not able to predict the future behavior of a customer on its own. It can only access past data and, for example, make comparisons with the customer’s behavior in previous years. Nevertheless, the analysis is suitable for deriving short-term actions (personal contact, discounts, etc.)
Catwing helps every online business to benefit from RFM(E) Analysis
Catwing is a flexible solution that offers RFM analysis tailored to various businesses. Our clients can execute full RFM analysis based on user-defined criteria, risk analysis, and schedule optimization. Learn more about other models and our pricing plans.