Last week, I explained what RFM (recency, frequency, and monetary) is and is not. This week, let’s discuss how we can actually implement it. Again, Jim Novo, author and former vice president of marketing and programming at Home Shopping Network, gave me tips on RFM’s applications.
Most direct and database marketing models are fundamentally about the allocation of resources to the highest and best use — in other words, return on investment (ROI) marketing. By understanding which customers are most likely to respond and have the highest future value, a company can allocate precious resources toward the customers representing highest value and away from those with low value. Jim says, “For each ‘discretionary dollar’ you might have to spend, it makes sense to spend it where you get the highest return. Most marketing programs involve discretionary dollars. You can do them or not, as you choose.”
In the simplest case, RFM is just about response. If you have your customers ranked by likelihood to respond, you can create much higher response for much lower cost. If you were to send two emails to the 50 percent of customers most likely to respond rather than emailing of all your customers once, you would drive a lot more response for the same cost. But, “do not use RFM scores as an excuse to inundate the most productive customers with promotions,” Jim cautions. “Like anything else, too much of a good thing is usually bad, and you will see your defection rate rise.”
You might sell several products or services from your home page. Another use of RFM is to take all the first-time buyers of these items, score them, and rank them by their RFM scores. You’ll see that certain items tend to generate first-time buyers who eventually have high RFM scores, and other items generate first-time buyers who eventually have low RFM scores. A high RFM score means a high likelihood to purchase again; a low RFM score means a low likelihood to purchase again. So, you eliminate or swap out the items generating first-time buyers with low RFM scores and allocate more space to products generating high RFM scores. You have just increased the efficiency of the home page, because the first-time buyers it now generates will average higher repeat purchases than in the past.
Say you are running two ads, and you have tagged the sources or created specific landing pages so you know where the visitors generated by these ads are coming from and going to. You score all the visitors from the two campaigns for recency and frequency with five divisions for each variable and get a set of 25 scores, from 55 down to 11. Visitors with high scores are the most likely to come back and have the highest future value. Those with low scores have a low likelihood of coming back and low future value. Ad A generates visitors with an average score of 45. Ad B generates visitors with an average score of 22. Which ad is better? All else being equal, ad A is better, because it generates visitors who tend to stick, relative to ad B. Assuming you want visitors to come back, this is a good thing. You shut down ad B and put your money into ad A, allocating the campaign resources to the uses with highest returns.
Search Traffic Quality
This is another version of the ad-targeting application, but it’s worth mentioning because so few people look at the quality of traffic they receive from search sources. If you care about these things, identify visitors/customers by search source, then score them by RFM — you can score visits, purchases, or both. Having different RFM scores for different activities is perfectly acceptable; in fact, it is highly desirable, if you think about it. If you find visitors from Google have an average RFM score of 433 on purchases and visitors from AOL have an average RFM score of 321, then the traffic from Google is more productive. You switch more of your advertising away from AOL and toward Google AdWords, because you know the quality of traffic coming from Google is superior to that from AOL.
The real power of RFM is in its predictive capabilities. If you can predict what customers will do, you can market to them more profitably. Nowhere does this make a pile of money faster than in customer retention programs. Though some think you should try turning low RFM scoring customers into high RFM scoring customers, I believe that’s a suboptimal use of RFM scoring. By the time customers reach a low score, “they have by definition a low likelihood to respond and low future value. Spending money on these already defected customers is not the highest and best use of capital,” Jim explains.
A better use of RFM for managing customer valuation is to use it to predict who is most likely to defect, so you can do something about it before the defection happens. This approach is particularly critical and immensely profitable when focused on high value customers — those with high RFM scores.
A recent report in The McKinsey Quarterly describes a two-year study of the attitudes of 1,200 households and stresses the importance of gaining a better appreciation of the underlying forces that influence customer loyalty — particularly their attitudes, changing needs, and buying patterns. This can help companies target efforts to correct any downward migration in people’s spending habits long before changes lead them to defect.
Most RFM scoring is done as a “snapshot,” a one-time event for a specific use — picking names for an email campaign or comparing the value of visitors between two ads, for example. But what if you turned RFM from a snapshot into a “movie”? What if you looked at RFM scores for the same customer over time and monitored these trends? You’d end up with a picture of a customer lifecycle, which you can use to predict and act on customer defection.
Suppose you have a group of customers who score 555 in the RFM model. These are the most recent, most frequent, highest value customers — today and in the future. They literally are the bread and butter of your business, the 10 percent who generate 90 percent of the profits. You do RFM scores monthly for your email campaigns, and you start to save the RFM score for each customer every month. You want to see if there are any trends — and there are! About 20 percent of your 555 RFM score customers drop out of the 555 position each month, moving down to a lower score. That’s a little troubling, for sure. The next month you see these same customers who dropped in RFM score last month drop even further in score. What you are seeing — literally before your eyes — is customer defection.
These highest value customers are marching off into customer oblivion, and you had better do something about it. The “something” depends on your business. Jim elaborates: “Trust me, saving these rocket-fuel customers is a lot more cost-effective than trying to reactivate the already low RFM scoring, already defected customers. The ROI on saving defecting best customers is absolutely enormous, with 400 percent return on investment an easy target. But you have to get to them during the process of defection, not after they have already defected. Have you ever re-signed up with the phone company when they call you three weeks after you already dropped it?” Do you think your customer would?
Marketers need to know what’s in their data and trim out the filler to provide continuous, data-driven ROI for their brands.
A new starter in Team SaleCycle recently asked me the following question… “Wouldn't they just come back anyway?”
American Apparel's chief digital officer discussed the future of retail, the importance of delivering value to the consumer, and strategies for an IoT and omnichannel world.
Every marketer has been sitting with his or her analytics team, reviewing an overwhelming spreadsheet of data points. It tends to hurt ... read more