Is the RFM Model Relevant Even Today?
Why do I need an RFM model (RFM Analysis)?
A key way to make your marketing effective is trying to understand the “value” of your individual customer. Termed ‘Customer Lifetime Value’ (CLV), there are many models that measure this metric. Developing a comprehensive and effective model of customer profitability requires an answer to the question – who are my most profitable customers? Is the RFM Model or RFM analysis relevant even today?
One of the models that have been in use for years to segment your customers to calculate their CLV is the Recency, Frequency, Monetary (RFM) Analysis statistical model. Using it for customer segmentation gives a picture of the past, showing what your customers were like, and is a good indicator of what your next goals should be.
RFM analysis is based on the Pareto principle, AKA the 80/20 rule. The rule states that ‘80% of the business comes from 20% of the clients’. Therefore, understanding who that 20% of your clients are is important, right?
RFM helps compartmentalize customers into clusters to identify those who are more likely to respond to promotions and also to upsell or cross-sell.
Interested in knowing how the RFM Analysis will work for your business? Click Here and our experts will get in touch with you.
This is how RFM Analysis works (in a nutshell):
Recency: Refers to the last time someone purchased from your business. What it means is a client who has bought recently is more likely to repeat as compared to one who hasn’t purchased for a long time.
Frequency: Refers to the number of times a client has purchased in a given period. The logic here is a customer who buys often will probably return in comparison to one who rarely purchases.
Monetary Value: Refers to the amount a client has spent in the same period. Obviously, one who has bought more is expected to return more often than one who has not.
The RFM analysis ranks each customer for each factor on a 1 to 5 scale (5 is highest). The 3 scores together are the RFM ‘cell’ for each customer ranking their historical propensity to buy with a ‘555’ customer ranking at the top. In this way, using RFM to segment customers, one can analyze each group to understand which one has the highest CLV.
But here’s the thing: Not all customers are created equal.
The RFM model (RFM analysis) may work for small and medium scale enterprises because of Its’:
- inherent simplicity
- effectiveness in direct marketing campaigns
- DIY nature
But if you are a large company having the wherewithal at your command, using RFM along with predictive analytics models is highly recommended. Predictive analytics does a far better job at forecasting sales and offers a better RoI segmentation based on RFM. But, then again, this is a costly method and not everyone can afford it.
What you must watch out for in RFM Model
You must be aware of the following:
There’s always the impulse to target customers with the highest rankings but that would be wrong.
Here’s why: It’s probable that customers who brought in the highest revenue may be casual. This information gives a clear signal that you should focus more on this group.
There’s also the instinct to ignore customers with low scores, and that too would be erroneous.
To avoid over-solicitation of high-ranking customers for it could lead to resentment making them flee from your business.
But the most important issue with the RFM model is the presumption that your best buyers will continue to be the best responders in your marketing campaigns. Yes, historical behavior does provide a roadmap for the future but it’s not truly predictive. As Anusham Acharya writes, this assumption ignores the fact that customer behavior might change over time or might have already changed.
New clients: where’s the frequency score?
Another issue for the RFM model is new clients. New clients tend to buy cheap products just to test the company’s services, which skews the model. But the other, more concerning problem is – how do you score new customers?
New customers have only bought once, so they can’t have a “good” frequency count, even though they may do well in the “Recency” and “Monetary” scores. So how does one account for them? Remember, the F in RFM stands for frequency, so clearly, that value cannot be derived for new clients.
To get around this hurdle, the data science team at Express Analytics has built a layer on the RFM Analysis to spot the potential among the new customers which helps in identifying those customers who are most likely to turn into high profile ones down the line.
Evolution of RFM Model
Since its inception over forty years ago, the RFM Analysis has evolved many times
Each iteration and variation involves mixing in new components to improve the model’s ability to predict.
Some examples: Ya-Yueh Shih and Chung-Yuan Liu (2003) proposed two-hybrid methods that exploited a weighted RFM-based method (WRFM-based method) or the preference-based Collaborative Filtering (CF) method to improve the quality of product recommendations. Their findings indicated that the proposed hybrid methods were superior to the other methods.
Rust and Verhoef (2005) provided a fully personalized model for optimizing multiple marketing interventions in the intermediate-term. This was done by conducting a longitudinal validation test to compare the performance of the model with segmentation models used in predicting the intermediate-term and customer-specific gross profit change. This battery of models tested included demographic model, RFM model, and finite mixture models. Their results show that the proposed model outperformed traditional models in predicting effectiveness in the intermediate-term (CRM).
Without a doubt, the RFM Analysis is a valuable marketing analysis and segmentation tool for many B2B businesses.
How well do you know your customers? Only when you know them inside out can you deliver the best customer experience. Our customer data platform Oyster helps you keep the Recency, Frequency, Monetary Value (RFM) score. With RFM analysis, your business can assess a customer’s propensity to buy. Want to more about how Oyster can keep the score for you?
An Engine that offers sophisticated RFM modeling algorithms
Liked This Article?
Gain more insights, case studies, information on our product, customer data platform