After the data mining … prospect research asks, “what then?”

I recently had a question from a prospect researcher who is taking on the task of learning data mining to predict propensity to make a major gift. (Yay!) She wanted to know, what happens “after” the data mining? Let’s say she ranks her prospects by score and now she’s got 100 or 200 names — what then? She writes: “I fear that I will then have to create 100 in-depth profiles on these prospects because the fundraisers will not have a plan or the confidence to move forward with these names.”

The situation is familiar: Too many names, not enough time to create full profiles for everyone on the list. My first instinct is to call this a prospect research problem and not a data mining problem.

When I was a prospect researcher, I had to create in-depth profiles for any prospect we were meeting with – even if it was the very first meeting and a gift would be years off, if it ever came at all. Today I work at a university with a much larger staff of development officers, but a research office that is (relatively) smaller. Full profiles for qualifying visits is unthinkable. DOs get no more than a summary briefing on prospects they’re meeting for the first time. This is for obvious practical reasons, but it’s my understanding that this is becoming the norm for many research shops – the full profile is produced only at an advanced stage of cultivation. So my first suggestion is, limit research to “top level” information only: Job title and company, giving history with the institution, maybe their Who’s Who profile if it exists … and not much more.

My second thought is that a data-related solution is possible. I would try an approach that Peter Wylie uses: Take the top several hundred prospects (that is, according to propensity score) and sort them in descending order by lifetime giving. Think of the propensity score as summing up the affinity that the prospect feels for the institution. The lifetime giving dollar amount also provides evidence of affinity, but capacity as well. If a prospect has a very high affinity score AND has given in five figures, they’re probably a good major-gift prospect. Take the top 10 or so names and do in-depth profiles on them alone, leaving the others for later. Or, if wealth screening data is available, one could use that instead of lifetime giving to cross with the list of top-scoring prospects.

But after thinking about it again, perhaps the real issue is contained in the original question: The researcher fears that fundraisers won’t have a plan, and they won’t have confidence in the process. That’s a fundamental problem, one that can only be addressed by communication, a certain amount of selling on the part of the data miner, and a lot of support from upper management.

Advertisement