Clarifying Your Project: A Way to Manage the Tyranny of the Urgent

When starting a data mining project, we are often challenged by our management to “go do some data mining.” We can get stuck from there, unless the tool that we’ll use has already been chosen, like a screening service or modeling vendor. Even then, making sure that we understand the question involved helps frame how we set up the study and then what we do with the results.

---

​For instance, “Find us more prospects” sounds like a clear directive; however, the related study can take many forms. This blog post considers the underlying questions related to this one request.

1. Your shop does not have enough major gifts prospects. For this scenario, find donors whose single gifts are the highest of your group, and model for prospects who look like them.

2. Your major gifts effort is waning. For this scenario, find donors whose giving is escalating in the current fiscal year, and model prospects who look like them.

3. Your annual giving program is not producing prospects for your major gifts program. For this scenario, look at the top of the class – prospects whose average giving during your most recent complete fiscal year, or on average over the previous three years, indicates that they are your top donors.

4. Your shop hired more gift officers who need prospects. For this scenario, look more closely at capacity rather than giving, or at philanthropy to other organizations, to find new prospects for in-person qualification.

5. Your shop is starting a campaign. For this scenario, look for your donors whose lifetime giving is highest and model prospects who look like them.

6. Recent economic changes have reduced your prospect base. For this scenario, look for prospects whose giving was good during poor economic times (2008-2012, for example).

The request, “Find more prospects,” can mean a variety of problem statements that are derived by understanding your business needs. Once you have established the underlying question, choose a good dependent variable for your study. We discussed dependent variables in this post from December 2015.

What are you studying? Tweet us your thoughts - @Staupell

This blog post was printed in the APRA Upstate New York newsletter in spring, 2009. Reprinted with permission.
​