RFM segmentation is a fundamental marketing tool within your CDP. It’s an intuitive way to divide up your customer base for marketing purposes. Using RFM segments can help you grow your business by organizing your customers according to their purchasing history and marketing to each segment in a relevant way.
The letters R, F, and M stand for Recency, Frequency, and Monetary value. Each of these three variables are measured in percentiles. Then we can denote each percentile with a corresponding ‘score’ from 1-10, including one decimal place, which is assigned to each customer.
R – Recency represents how many days it’s been since the customer’s most recent purchase. R scores closest to 10 are assigned to your most recent customers.
F – Frequency represents how often the customer makes purchases. It’s calculated for each customer individually, within a one-year look-back window.
M – Monetary value represents the total value of all purchases made by the customer. The percentiles for monetary value are calculated relative to the entire customer base.
Measuring RFM in percentiles allows you to keep your RFM model relevant no matter how the data set changes. Over time as you take advantage of your RFM segments, the frequency of purchases and monetary values associated with your top 10% could improve significantly. Then, the values that were initially associated with your top 10% would only be mediocre. You want to avoid guiding your marketing strategy by an unmoving barometer.
E-commerce platforms like Shopify or Magento have all the necessary data to create an RFM model. But they don’t necessarily have the tools to develop the model itself or do the segmentation. BigCommerce uses OroCRM as an example of how to create RFM segments from BigCommerce data. The analytics platform Putler also suggests using excel or python and their own platform for RFM analysis.
Kevel uses data from your e-commerce platform as well as first-party data collected directly. The RFM model is built-in and runs automatically. So you actually skip past the analysis part right to the segment creation part.
Segments are created using ranges of scores. So a segment using M 9-10 represents the top 10% of your customers in terms of how much they each spent.
Here’s how it looks:
Caution: Avoid over-complicating.
Remember, this is meant to be the easy, intuitive method of segmenting customers. Some recommend using quintiles or quartiles (splitting your audience into 5 or 4 sections) as a way to control the number of potential segments. And sometimes, those people even have use cases for every segment within their imposed limit.
Our recommendation: Don’t sacrifice granularity, but don’t make segments you’re not going to use, either. Just because you CAN make 1000 segments doesn’t mean you should. We start with only three: High Spenders, Best Customers, and Churned Customers.
This segment captures your highest-spending customers using Monetary value, regardless of their Recency and Frequency scores. There are two ways we recommend using this audience.
This segment captures customers who spend a lot, have made frequent repeat purchases, and have made their latest purchase fairly recently. They are active buyers who have made purchases recently and will probably come back again.
You can use this segment both for retention and acquisition, just like the high spenders audience. But since this group is your most loyal cohort, It’s worth it to reward them with special offers or exclusive deals.
These customers have made purchases in the past but haven’t been back in a long time. There are many potential reasons for a customer to churn. But there’s always a possibility that you could win them back with special offers or by rectifying a poor experience. Retaining customers is, after all, way more cost-efficient than acquiring new ones.
The three examples above are default segments that come pre-configured for all of our clients. But it’s easy to duplicate them and customize them as needed.
For example, it may be relevant for growing businesses to designate the top 30% of customers as the best customers instead of the top 10% if the volume is not yet enough to be meaningful.
Once you have your fundamentals, there’s no reason you can’t play around with additional parameters like the mad marketing genius you are. Here are a couple of ideas.
A common complaint with RFM segments is that they don’t account for how long a customer has been with you. Loyal, long-term customers are not necessarily your highest spenders if they usually make smaller purchases. And they are not necessarily your most frequent customers either. But you still want to reward them for their loyalty and incentivize them to keep coming back.
To capture this segment, set up your RFM segment with middle-range scores. Then add a rule so that the first purchase for each customer was made before a specific date.
Now and then, you might pick up trends in your analytics showing outstanding performance with specific locations or demographic groups. When that happens, you have an opportunity to capitalize on that trend by creating an RFM segment just for this cohort. Double down on your success by using the high spenders audience with the lookalike trick we described above. That way, you make sure you attract the top-spending customers in the cohort.
Starting from scratch to create RFM segments is math-heavy. So if you’re not the type who wants to spend time with python or excel sheets, you’ll want to consider a software solution. Even if you are the analytical type who loves crunching numbers, getting an automated solution can save you tons of time and remove the bother of keeping the models updated.
Ideally, you want something that incorporates new data automatically, so your data set never gets old. Getting a solution (like Kevel Audience) that automates segment creation and integrates with ad platforms further eliminates the gap between analysis and activation.
The best way to learn anything is by doing. Want to see precisely how RFM segments can help grow your business? Talk to us today about starting a trial.