You probably know this winner!

Buzz words abound in the marketing industry.  “Modeling” is one that has recently been grabbing my attention.  I hear it being sold in many impressive-sounding ways.  But the base concept is to find relationships in data that predict meaningful, repeatable outcomes.

As marketing budgets tighten, marketers are increasingly looking for ways to improve the success of retention and reactivation strategies in order to minimize the burden of necessary, albeit costly, acquisition programs. Exploring whether or not modeling can save the day seems to be a fair-enough approach.

Echoing on that theme, at their recent international convention in San Francisco, the Direct Marketing Association (DMA) challenged the industry to develop best-of-class modeling for nonprofit donor reactivation. Sixteen companies, with teams as large as 15 members, utilized more than $1 million worth of the latest, big-name analytics software to develop predictive models from more than 2,000 distinct variables.

After hours of bagging, bootstrapping, binning, clustering and regressing (as I said earlier — impressive sounding), who was the winner?  What were the three variables that, by far, outperformed all others?  R, F and M.

Long hailed as the gold standard in fundraising circles, Recency of last gift, Frequency of giving and the Monetary value of gifts proved once again to be the greatest predictors for cost-effectively  reactivating nonprofit donors.

So what does this mean for the nonprofit organization that is somewhere in the process of investing tens of thousands of dollars in exploring the potential of modeling for improving stewardship of their marketing dollars? — Be informed. The industry has yet to consistently produce an alternative that doesn’t build off the fundamentals of RFM. Modeling can certainly provide incremental gain; just make sure the cost and return of the model is justified over the baseline of your well-honed, RFM-based system.

Testing is good.  Innovation is great.  But following the “buzz” sometimes means you get “stung.”

Michael Fassnacht

Innovation Services Director of Analysis