Note from ProspectResearch.com: Nicole Bechard is a Product Manager for Modeling Products at Target Analytics. You may reach her at firstname.lastname@example.org.
Recently the attention of the entire country was captured by a machine, in fact a very smart super computer named Watson, and its deep analytics capabilities. As a self-proclaimed analytics geek, I quickly marked the Watson Jeopardy challenge on my calendar and waited in anticipation of the next man vs. machine challenge from IBM. I was quite intrigued as to how a room full of computers could unravel the Jeopardy wordplay in order to come to the right answer.
We watched in amazement as Watson answered question after question correctly and generally displayed the right answer on the rare occasion one of the human players had the chance to ring in first. I found myself awe inspired by the amount of data it could quickly analyze. Although the human players couldn’t beat the machine’s speed, as Ken Jennings graciously articulated in a recent article for Slate, this was a win for humankind in the end. The contest demonstrated that our ability to solve problems with analytics has risen to a new height.
You may be saying to yourself that as impressive as Watson’s performance was, analysis at that level is far too complicated to gain any benefits unless you’re among the most statistically-literate individuals. Let’s look at this from another perspective. The true benefit for those of us without statistics degrees isn’t the algorithm itself but the ability to sort through millions of pieces of data to arrive at a simple rank. In fact all predictive models result in a score, ranking or classification of some kind; regardless of how complex the methodology the result should always be very simple to use.
Analytics is just a framework for logically looking at many data points together and predictive modeling is a way to uncover patterns in that data. Even if we don’t have insight into how the Watson algorithm was built, we can still appreciate the takeaways of analytics and how it easily leads to better decision making. This curiosity is the first step in discovering how modeling can reinvigorate your giving programs.
Frequently I hear fundraisers and prospect researchers say that predictive modeling is too advanced for their organization and that one day they hope to be savvy enough to understand the results of a modeling project. I can’t stress enough that you don’t need to hold a PhD in statistics or be a computer programmer at IBM to effectively use model scores in your everyday fundraising practices. Predictive models can serve a variety of objectives and be tailored to fit any sized organization. These analytic tools aren’t reserved for only those who can comprehend the mathematics behind them; they’re a good fit for any researcher who wants to take their prospecting beyond one or two points of information.
If you still don’t think you’re ready, ask yourself these questions:
- Are you being asked to raise more money with fewer resources?
- Do you find that you have pools of identical-seeming donors that you can’t differentiate from each other?
- Do you find your rates of retention and upgrade on the decline despite your best efforts to counter this trend?
- Do you feel like you’re going back to the same donor pool over and over again?
- Are you overwhelmed with a large database, not knowing where to start?
- Do you make data-driven decisions when assessing your prospect’s commitment and capacity?
- Do you find that you don’t know which of your wealthy donors are most likely to commit to a large gift?
If you answered yes to any of these questions, your organization could benefit from implementing predictive models. Employing a simple set of scores will allow you to make quicker, more informed decisions about your donors and prospects. What is required from you, however, is a willingness to learn what the scores mean, listen in an unbiased way to what they’re telling you about your best potential donors and change your solicitation strategy to use the results to their fullest capacity