Missy Garner, Former Director of Prospect Research
Rick Loveday, Prospect Research Analyst

Part 3: Refining the Model

This post is the third in a four-part series that will outline how Clemson University’s Prospect Research Team developed a philanthropic model score.

Part 1: How to Develop an Analytics Project: Building a Model Score

Part 2: Building the Model

We built our model score using a whole number scale of 0-5, with 5 being the best. As we applied this model score, we realized it did not differentiate between the different levels of giving; it merely stated whether an individual hit the minimum requirements. We had no idea if an individual gave $15,000 or $100,000 to political causes. We also could not distinguish between a person who gave a $5,000 philanthropic gift with one who gave a $500,000 gift. We quickly came to the realization that we had to develop the model further to account for the wide range of gifts. In order to solve this, we decided to incorporate a decimal-based scoring system.

The concept behind the whole number score worked well to highlight which groups a constituent fell into. Could that same concept be used on the decimal side to highlight which giving ranges that same constituent met? We decided to use the numbers 3, 6, and 9 for our giving breakdown. The number 3 meant that the constituent gave enough to meet the minimum threshold, which was $15,000 in lifetime political giving or a $5,000 single philanthropic gift. The 6 was applied to those with over $50,000 in giving, whether in lifetime political giving or in a single philanthropic gift. The 9 represented over $100,000 in giving. Since there were two types of giving, we decided to use two decimal places. The first place would represent the philanthropic giving while the second decimal would represent political giving.

When we applied this tweak to the score, it proved, almost immediately, to be a significant upgrade over the first model. We quickly realized some of the major deficiencies in the original model. Using the whole number score, we thought that the more categories a constituent was in, the better that person was as a prospect. The tweaked score immediately proved that to be false.

With the whole number method, a constituent that appeared in all five categories had a score of 5. Another constituent that only met the philanthropic and political giving categories but missed the non-profit/foundation association, the real estate ownership and the SEC insider/business executive aspects had a score of 2. So, according to the whole number score, 5 is always better than 2.

When we applied the decimal score to those same constituents, the constituent with the 5 only met the minimum giving levels and had a new score of 5.33. The constituent that had the score of 2 had a largest single philanthropic gift of $50,000 and a lifetime political giving total over $100,000. This updated that constituent’s score from a 2 to a 2.69.

In looking at the updated scores, we could make the argument that a score of 2.69 is better than a score of 5.33. While both constituents are acting in a philanthropic manner, the constituent with the 2.69 is giving at a significantly higher level and thus is potentially a higher priority.

As a result of this change, the decimal places became the primary focus and the whole numbers became a minor piece of extra information. We found a whole new crop of potential prospects to focus on. Our next task was figuring out how to introduce the model score to our Development Officers and explain how it would be useful. We expected this to be a more daunting challenge than actually creating the model score since our fundraisers lived in a world where hard assets, such as real estate values and salaries, were the basis for determining which constituents they would pursue. We were about to ask them to potentially ignore a constituent who is the president of a company and lives in a $4M home - but had given nothing philanthropically or politically - and instead focus on a constituent who owns a small business and lives in a $200,000 home, but who has donated a significant amount of money to various organizations.

The next, and final, blog post will detail how we introduced our model score to our Development Officers.

Comments have been disabled on the Apra-Carolinas blog. Please visit Rick's post on LinkedIn and share your thoughts there!