We Created The First-Ever AI Model to Identify Advocates to Acquire. And it Worked.
By Alex Stein, VP
When we launched AdvantageAI, we knew two things: 1) it would help mission-aligned nonprofits and campaigns find prospects and donors, and 2) serve as a laboratory of innovation. From that laboratory, we’ve already identified one new offering that has the potential to generate transformational results for movements focused on advocacy. In one case, it already has (more on that in a moment).
AdvantageAI is set up to help nonprofits and political campaigns solve today’s problems, yes, but as important is what it can yet be. This first-of-its-kind data co-op is characterized by an infrastructure made to evolve:
- Members are from more than 60 nonprofits and campaigns, and AdvantageAI is constantly growing.
- There are more than 70 million unique email addresses, all with “data stories” our algorithms use to identify the causes for which they’re most likely to align.
- The combination of machine learning and unparalleled insight into the needs of its members adds up to a service that’s always getting better.
With that machinery at our disposal, when an organization asked us if we could create a model to identify new supporters who would call Congress to advocate for passage of historic legislation, we said yes – and the results exceeded even our own high expectations. We helped the member generate calls to members of Congress at what was – for them – a record-low cost-per-call.
To be sure, this was just our first toe-dip into using our massive data lake and machine learning techniques to do advocacy modeling. But the results were encouraging, and even in isolation, we set up a test to evaluate the genuine effectiveness of the approach.
For starters, we compared the advocacy model to a traditional model, which, in our case, is used primarily to identify prospective donors. The advocacy model was 21-percent more effective at driving calls, statistically significant at the 0.05 confidence level.
Again, there are things we still need to know about using AdvantageAI in this way. But if this holds – if what we’re seeing continues to prove out – we and our partners are on the cusp of something potentially transformative for identifying and compelling advocacy actions.
We’re all in to find out. If you are, too, shoot me a note at email@example.com and let’s learn together.