Your recency, frequency, monetary value (RFM) model can tell you a lot about how your business is doing. Assign value to each category, and look at the resulting matrix, then you can use that data to target customers and create various slices of your customer base to test and market to. In fact, testing your RFM hypotheses can be extremely valuable to your long-term success, if you’re doing it right.
Every model begins with good data. In order to use your RFM model, you’ll need to first make sure the data you’re mining is reliable. If you use a CRM, work with your agency or team to give your data a check-up. Ask questions like: Are we collecting the data correctly? Check for anomalies in your database to make sure the data is on target. How reliable and accurate is our scoring method – if you haven’t updated it in a few years, perhaps it’s time to change the values for your scoring based on market fluctuations or differences in how your selling today vs a few years ago. Is it time to purge some customers who aren’t coming back to the brand? Have we thoroughly eliminated old data? Creating segment-based testing. As you develop a more accurate database to work with, you can pull together your matrix.
Create your RFM matrix based on the data and statistics that are relevant to your business. What is “recency” in your market segment? If you’re a big box retailer, the numbers will be much different from a luxury brand. Frequency can also have various values – there are frequent buyers and frequent browsers, and both may be very important to your business. Finally, look at customer spending wisely, is this a young customer who is just growing into your brand? You don’t want to discount potential customers. Once you’ve established the metrics that are right for your business, you can create a matrix of your segments, and begin looking at ways to test each segment.
Testing based on RFM models
Use your RFM model as part of a broader strategy to segment and target customers. As CRM Trends points out, there are some factors that RFM doesn’t take into account. For example, the data won’t provide a reason for low-scoring customers’ behavior, but RFM does gives you an easy way to segment, and helps you simplify this complicated process. Look at your matrix to begin coming up with ways to cultivate high-scoring customers more, and to bring low-scoring customers back into the fold. Don’t neglect the middle – figure out what you’re doing right for them, and start testing ways to move a percentage of the mean into the top layer. Develop tests to see what it takes to gain customer’s interest, and as you see success with programs in different target areas, broaden your scope to capitalize on the success of each positive test.
The book, High Performance Marketing, notes that RFM can be very effective way to begin using your data to optimize your programs. One executive interviewed for the book noted that RFM may miss some nuances like price sensitivity, “but the rush to get stuff out the door and the daily operational pressures make it very difficult for us to invest a great deal in testing and R&D into these issues.” The RFM model meets the needs of companies that are moving fast and helps them begin to clear through the clutter.
As you develop a better grasp of your sales trends through RFM modeling, you can take the next step toward more sophisticated mathematical models for digital marketing. This important first step is a lower-cost option that forces you and your team or agency to define your audience and your goals, and helps you test hypotheses and assumptions so that you can start seeing the value of predictive modeling. As you grow to love the data you’ll start to look at more sophisticated solutions.
