Focus less on the model,
more on the mindset

woman-looking-at-art_855x550

By Kathy Howrigan, Senior Consultant & Principal, Analytical Solutions; and Alison Rane, Associate Consultant, Analytical Solutions; Manager, Research & Development

Analytics, data mining, modeling – the fundraising world has been awash in these terms for years, and they show no sign of abating. But is analytical thinking really becoming part of the culture of fundraising? Or has focus on particular techniques boxed in our understanding of how data can support the business of fundraising?

Rather than simply a toolbox of statistical techniques, fundraising analytics should be considered a larger process:

  1. Input: raw data and defining questions about, or goals for, your institution.
  2. Analysis: exploring data to test hypotheses using statistics.
  3. Output: not a model’s output; instead, using analytical evidence to support changes in your business processes.

Too often the focus is on Analysis (#2), and the process stops short of Output (#3). Predictive models and scoring have, to some degree, come to define fundraising analytics. But analysts should be empowered to think beyond tinkering with technique and toward a holistic, experiment-driven view of the core business processes of fundraising:

  • Get a handle on what has been happening at your organization in recent years, beyond total revenue. Run your reports (or query your database) and compare results to the fundraising activities over that period. Slice that data in different ways and identify the story of your business over the past five years from both the perspective of fundraising activities and dollar/donor results.
  • Determine what questions this examination raises about your business processes. Did a particular push for Parent giving fall flat last year? Through which giving channels did you see better results? How would you describe the appeals (or the user experience) on those channels?
  • Test, test, test. Encourage and embody a spirit of experimentalism. Use previous giving to assign aggressive ask amounts to one random population and conservative ask amounts to another (building testing cohorts is a great way to use those predictive scores).
  • Evaluate the results from your tests against your expectations and current business processes. Present findings to decision-makers and work with them to use data to determine “what’s next” for your fundraising operation. That’s analytics!