A long, long time ago (i.e., 2006) in a geocode minutes-to-hours away (i.e., MIT), the Powers-That-Be in Resource Development hired an ABD (i.e., All But Dissertation) to continue the work done by a consulting firm in building predictive models. So, for two years, this “padawan” data mining and modeling specialist plugged away, building major gift donor models, analyzing planned gift donor data to create time-to-(unfortunate but inevitable) event models, and providing statistics for whatever projects may warrant such number crunching. Alas, the enticing but dark lure of Corporate America proved too strong for the newly minted degree holder. Thus, for two more years, he toiled in the healthcare market research space, learning the dark, SASsy art of data and proc steps and weaving patient and healthcare provider surveys along the way.
Like many stories of redemption, however, a new hope arose in the opportunity to build a development analytics program back in the non-profit space, this time at Dana-Farber Cancer Institute (DFCI). Under the tutorage of research master Barbara Moore, the just-knighted assistant director first evangelized the concept of predictive analytics to various business units within DFCI Development and the Jimmy Fund. At the same time, he reassured the Dev IS database group leader of his benevolence towards the database and non-infringement on IS job functions as he learned more and more about the enormity and complexity of the databases. During that inaugural year when he wasn’t navigating the politics or learning the ropes, the assistant director was able to develop predictive models for business units dealing with principal/major gifts, annual giving, gift planning, as well as signature events for the Jimmy Fund such as the Jimmy Fund Walk and the RadioTelethon. In the ensuing years, both the apprentice/assistant director and his research master explored the utility of predictive models and expanding analytics services. This was accomplished either by increasing the number of projects for some business units such as Annual Giving or by changed project directions from developing predictive models to something more descriptive and exploratory such as segmentation of current target populations, in order to uncover potentially disparate subgroups for differential marketing. Over time, team members within the various business units saw the assistant director, along with the entire Research group as a whole, in a more business consultant and strategist light, and less and less as backroom analysts or walking statistics software.
‘But where to go from here?’ questioned the assistant director, sensing a nagging disturbance in the force developing out of the business analytics realm. The pull, as it turns out, is not of the corporate darkside but rather from tales of attaining the avatar status of the Data Scientist, a being able to acquire, merge and transform structured and unstructured data at will, to use supervised and unsupervised statistical algorithms to elucidate data trends, patterns, and target group suspects at the drop of an open-source hack-a-thon, and, as the same time, have the ability to enlighten her or his clients with clear, stimulating presentations and bold, new, successful strategies.” <cue dramatic orchestral music and fade to black>
Besides trying NOT to bore you with the verbal equivalent of “my vacation: a slideshow,” I want to point out a couple of epiphanies that I recently had about building development analytics programs. Perhaps obvious to everyone else but me, the first “eureka” is that no “one right way” exists for creating and building development analytics programs. My story is just one way in which research shops have incorporated analytics into development. Some development analytics programs are nurtured by focusing on one or two particular areas of fundraising such as principal/major gifts and gift planning. Another approach may be to cast the net far and wide, to see how analytics may be applied to many and varied business problems. Furthermore, development analytics programs do not all have to rely on the same techniques and technologies either. While my background includes learning some predictive modeling techniques, as well as developing people skills and public speaking, it does not include learning languages such as SQL or Python or R. Thus, the development analytics program at Dana-Farber builds many models and act as another consulting/strategizing voice to its internal clients; data retrieval must come from Dev IS.
© 2018 New England Development Research Association