This article was first published as “Marianne’s Mining Minute” in the APRA Upstate New York chapter’s newsletter, 2009.
What It Is
Recency, frequency, and monetary value is a collection of measurements used by data mining firms to assess a customer’s potential. Called RFM, this tool looks for a company’s highest value clients.
How We Differ from For-Profit Firms
In fund-raising, we have to remember that our prices, unlike, say, Coca Cola, are widely varied. To buy a Coke, I might pay $1.00 for a 12-ounce can, or as much as $10.00 for a case (my Coke purchase prices may be way out of date, my apologies). In fund-raising, however, our price varies from a $25 starting annual fund gift to a $500 million top level campaign gift. We can not measure donor behavior in the same way that Coke measures our consumption.
The other concept to keep in mind is that major gifts prospects make pledges rather than give cash gifts. The old adage that major gifts come from assets and annual fund gifts come from the checkbook/credit card still holds for most gifts. A major gift donor makes a multi-year pledge.
Managing the Interval Problem
So, a major gift is going to show a gap in the recency and frequency measurements, if the downloaded gift data shows pledges and outright gifts only. If the gift data, instead, shows pledge payments and outright gifts, then the recency and frequency scores would improve.
Annual fund donors, however, can give by credit card every month. That leaves their recency and frequency scores much higher than major gifts donors who are making annual pledge payments with occasional annual fund or special purpose gifts.
Using RFM, Anyway
One method I would suggest for using this tool is to both measure it against single pledge payments and outright gifts, making sure that the payment dates are included. This way, one can tell if a higher RFM means an annual fund gift or a major gift.
I would also encourage the analyst to separate each score. Would a higher frequency score mean a lower level gift? My studies on this question were always yes, but it has been several years since I tested it. I have also seen the effect of giveaways (a la public television and radio), where the recency and frequency scores were high for donors whose giving levels stopped at the highest reward giveaway level.
Finally, it could be true that RFM is low for major gift donors, which could be used as an inverse indicator of a donor’s likeliness to become a major gift donor in the future. What if it works?