Table of Content
User segmentation is critical for every mobile marketer to better understand users, satisfy them, and develop more personalised products and services.
That being said, all successful businesses today, utilise different techniques to identify customer behaviours, requirements and needs – as sustaining loyalty is a key task for marketers, sales teams, and business owners.
In the quest to build better relationships with customers, finding a lasting competitive advantage exploiting the tons of customer data available is a smart move nowadays. Fortunately, one of the most effective analytical techniques to exploit and transform this data into meaningful knowledge is RFM Analysis.
So, What is RFM analysis?
The RFM (recency, frequency and monetary) analysis model is predominantly a behaviour-based model utilised to analyse the behaviour of users/customers to make accurate predictions of their actions.
As a proven database marketing technique, RFM has traditionally been employed by catalogers to dramatically increase conversion rates and reduce expensive costs of mailing catalogues. However, nowadays, many online retailers use RFM analysis to primarily augment conversion rates, segmentation, personalisation, relevancy and revenue.
Characteristics of RFM Analysis
In practice, RFM analysis is basically a three-dimensional way of classifying, or ranking, users/customers to determine the top 20 percent, or best, customers. Fundamentally, it’s based on the 80/20 principle that 20 percent of customers bring in 80 percent of revenue.
In direct marketing, RFM analysis is based on simple theories of:
- The most vital factor in identifying users who are most likely to respond to a new offer is recency. Customers who buy more recently are more likely to make a repeated purchase than customers who purchased further in the past.
- The second key factor is frequency. This implies that customers/users who made more purchases in the past are more likely to respond than are those who made fewer purchases.
- The third most crucial factor is the total amount spent (monetary). This theorizes that customers who spent more (in total for all purchases) in the past are more likely to respond than those who spent less.
All things considered, RFM analysis enables mobile marketers to find answers to the following questions:
- Who are your best customers?
- Which of your customers can be retained?
- Which customers could contribute to your churn rate?
- Who has realistic the potential to become a valuable customer?
- Which of your customers are most likely to respond to engagement campaigns?
How Does RFM Analysis Really Work?
The RFM analytic model was first proposed by Hughes, Strategic database marketing (1994) to represent customers’ consumption behaviours based on the transactions in a computerised database, broken down into three key variables:
This variable relates to the number of days since the customer’s last purchase. In practice, it focuses on the period of time starting from when recent consumption behaviour occurred (last purchasing) to the present purchasing.
As such, the closer the date to the present, the higher the possibility customers will make another purchase. Hence, a higher value in the recency variable. Usually, this attribute’s value is defined in days. For instance, if the customer’s order was 40 days ago, then their recency input is “40.”
This variable refers to the number of orders placed in a specific time period. It is expected that the higher the purchase frequency of customers, the higher loyalty customers are, and the higher the customer’s value to the business. As such, if a customer placed six orders over the course of one year, then their Frequency input is “6.”
This variable revolves around the total amount of money spent by the customer over a specific time period. It’s expected that the higher the monetary value, the higher the profit contributions made by the customer to the business, and the higher the customer’s value. For instance, if a customer has made 5 orders of $50 each, over the course of 1 year, their monetary value input for the year essentially is $250.
In summary, the objective of RFM Analysis is to essentially segment users/customers based on purchase behaviour by understanding the historical actions of individual users/customers for each RFM factor. In practice, customers are ranked based on each individual RFM factor, and all the factors are combined to create RFM segments for targeted marketing.
Why Use RFM Analysis?
RFM analysis is a valuable marketing tool to improve financial results and client satisfaction during direct marketing campaigns, which positively impacts the way businesses relate to their customers.
In practice, by precisely estimating the likelihood of each client making a purchase, RFM optimised models increase the financial results of marketing campaigns, help to identify a business’ most valuable clients and improves the quality of communication between organisations and their customers.
With better insight into customer behaviour, all these benefits help companies to become more efficient, competitive and profitable, and thus stay in business for a longer time.
The positive impact of RFM modelling is both qualitative and quantitative in nature. Its qualitative nature stems from the fact that the optimised models motivate businesses to send marketing communication only to clients that are likely to make purchases. And as a result, such customers will appreciably value the offers made to them.
On the other hand, the quantitative nature is because of the cost reduction it provides by preventing companies from wasting money on customers that are unlikely to make any purchase at all. This money saved on ineffective communication can be used in new marketing campaigns that offer products or services better suited to the customers they target.
The Advantages of the RFM Model
The RFM model has been popular in direct marketing segmentation for the last two decades for several reasons:
- It’s a cost-effective way of acquiring actionable customer behaviour analysis and is relatively easy to quantify customer behaviour. This is because customers and transactional data can be stored in an accessible electronic form and exploited in different ways. As such, RFM variables are gathered via an internal database containing customer-specific information regarding the transaction history and are not obtained through the aggregate level information in the demographic databases.
- The RFM model is very valuable in predicting response and can help boost a company’s profits in the short term.
- Purchase behaviour can be summarised utilising a very small number of variables.
- RFM analysis is more meaningful for targeting particular customers.
- RFM analysis can be used to measure the strength of customer relationships as it can effectively identify valuable customers.
The Disadvantages of the RFM Model
Despite being vital for developing marketing strategies, the RFM model has several disadvantages.
- Given that RFM seeks to primarily identify valuable customers, it only focuses on the best customers. As a result, it provides limited meaningful scoring on recency, frequency and monetary when most customers don’t buy often, spend little and have not purchased lately.
- It also ignores the analysis on new companies setting up in a short period and customers that only purchase once and place small orders.
- The RFM model can only utilise a limited number of selection variables, yet most household characteristics have an effect on the probability of customer response.
- Unless RFM-variables are all mutually independent, the RFM analysis model does not double count.
- RFM analysis tends to focus on a company’s current customers, and thus, cannot be applied to the prospecting for new customers as the marketer does not have transactions for prospects.
- The RFM model estimates a single response model for all users/customers in the database, and hence, assumes the homogeneous customer database, which is usually contrary to the real situation that users/customers typically have a considerable heterogeneity.
- Lastly, the RFM model isn’t predicted as a precise quantitative model, and the importance of each RFM measure is different amongst disparate industries.
Main Steps for RFM Analysis
There are three basic steps involved in RFM analysis:
- Sorting all users/customers in ascending order based on Recency, Frequency and Monetary Value.
- Splitting customers into quartiles for each factor.
- Combining the factors to group users/customers into RFM segments for targeted marketing.
Generally, the first step entails identifying and sorting customers by downloading a spreadsheet with customer purchase history. Ideally, this file should include the date of the most recent order, the number of orders placed over a selected time period, the total value of all purchases made in that time period, and user/customer IDs. If possible, it should also include the customers’ email addresses.
Once your spreadsheet is ready to be segmented, sort each column in ascending order based on its RFM factor. For instance, select the column for recency, then sort it, so the most recent orders appear first, and the oldest orders are last (named as R=1 or R=2)
In the next step, score your customers using basic ranking, as it will serve the needs of most mobile marketers, but be aware that other approaches do exist, and might be more appropriate for your business.
With users/customers now organised in ascending order, divide them into quartiles, or four equal groups for each designated RFM factor. In practice, the customers in the top quartile should represent your best customers for each factor. For instance, the top quartile for Monetary Value should have the 25% of your users/customers who have spent the most.
With customers in quartiles, you can proceed to group them into RFM segments. For example, if a customer purchased a particular item 17 days ago (R=1), bought seven times in the current year (F=1), and spent $578 total in the current year (M=1). As a result, you will place this customer in RFM segment “111” that contains your “Best Customers.
As has been noted, RFM analysis is essentially an effective method used to identify high-response customers in marketing promotions, and improve overall response rates.
Fundamentally, while communicating with your users/customers, taking a one-size-fits-all approach is not really helpful. So, it’s important to segment them based on demographics, brand affinity, psychographics, spending capacity, brand affinity, and so on.
Overall, understanding appropriate segments to target necessitates an intuitive understanding of your customers, and RFM Analysis takes a data-driven approach to segmentation by considering:
- How recently did the user make a purchase?
- How much did they spend?
- How frequently did they make purchases?