From Buttons to Blogs

With so much intuitive data unavailable to the computer, any “decision” (or predictive model) is likely to be wildly distorted. So then, what we need to do is adjust our expectations.

Instead of expecting software to give us answers, we should set it to the task of alerting us when answers are needed: “Calls to new donors have dropped 25 percent since last week” or, “Five major-gifts proposal deadlines have slipped.”

A simple decision-support system can be developed to flag these conditions quickly and alert human resources. But by the time calls to new donors drop by 25 percent, it’s probably too late. Maybe what we really need is an “early warning” mechanism.

Consider: You target a donor for a major gift and set up a proposal in your data system. If you have a history with this donor, you might be able to calculate a ratio of “number of phone calls to gifts received” or the average time from the start of cultivation to the receipt of a gift. You could keep your IT staff busy making those calculations and building models based on even more sophisticated analysis — all of which will look impressive.

But then how do you avoid the “deadline shuffle” that keeps the same six prospects on a “to be asked in the next six months” list for more than a year? A good decision-support system can keep us honest and help us make the most of “what-if” thinking: “What if the deadlines on my A list slip by three months?” “What if this appeal is late?”

In his book, The Haystack Syndrome, Eliyahu Goldratt asks whether an information system can answer the question (not project, but answer): “How much net profit will our company make next quarter?” His answer: Yes.

Might we expect the same from our fundraising systems? I suggest that we can, if we follow Goldratt’s methods. This means: the rigorous scheduling of gift “production” activities (scheduling backwards from the deadline, you define a series of “moves” to advance the gift); development of mechanisms to alert managers when gift-production activities start late (or fail to start at all); and scaling back or abandoning analytical efforts to spot trends or profile donors, and relying instead on the intuitive data possessed by our best fundraisers.