Why AI Modeling is the Future of Acquisition
What makes the traditional co-op model so powerful? And what happens when we take it a step further and apply cutting-edge AI to deliver high-quality and dependable leads to members of the first-ever online co-op?
I sat down with our own Chief Product Officer, Michael Milton, for a Q&A covering the basics of our AdvantageAI co-op, data modeling, and how it can all work for you.
MH: Let’s start with the basics: What does it mean when we say we “brought the direct mail co-op online?”
MM: A co-op is a platform that aggregates data about how people are interacting with member organizations’ content. Because we’ve brought that co-op model online, that interaction data refers to clicks, opens – any way someone engages with emails. And what having that aggregated data means is that we can do really advanced things with modeling. We can look at our “data lake” of email addresses and make predictions about a variety of factors such as the likelihood of each individual to donate to a particular organization.
MH: So what makes this online co-op more powerful than data analysis I might be able to do through a CRM?
MM: We’re working with terabytes of data from partners’ CRMs, creating a massive data set that, frankly, is too big and unwieldy for the CRMs to get a lot of use out of by themselves. Complex analytic queries on CRM data sets are usually either not permitted, or run the risk of destabiizing the CRM’s infrastructure. With our data lake architecture we can interactively query data no matter how big the data set is and never have to worry about crashing a CRM. There is no question too complex for us to ask of the data.
MH: What does this look like in action?
MM: What we’re fundamentally doing is watching the patterns of behavior individuals are expressing over time and trying to understand those patterns, and then build audiences based on those patterns. And then we extract as much information out of that data as possible.
To put that all into practice, what our reactivation model can do is – let’s say Melanie signed up to the list six months ago, and she opened and engaged in a certain pattern. We know from other analysis that she was behaving in the same pattern as our other strong names in this pool. So even if she dropped off for a period of time for whatever reason and then we suppressed her after that, we still know from that pattern that she’s a strong candidate for reactivation.
MH: I’m following how this works, but how is this different from other fixed-cost acquisition?
MM: To say it simply, we have vastly more data. And not just more, but more diverse, more scalable, and more predictable outcomes. A petition-style swap may give you a good chunk of new names, but they can’t tell you the signers’ propensity to donate again. Our co-op can.
Another way to look at this is that through AdvantageAI, we are helping nonprofits connect with people primed to support them. So we consistently see from these names higher open rates, lower unsubscribe rates – conversions that suggest we’re providing nonprofits with information they didn’t used to have about who really wants to be part of their mission.
MH: Is there anything else we should know about the AdvantageAI co-op and its future?
MM: We always believed this platform would drive outstanding results for members, even before we began creating it. But of course, having intuition that something will work doesn’t guarantee it, or tell you how, or give you a sense of how much and how well it will work. What we’re seeing in the actual results as more and more groups join the co-op and build out that data lake is a very tangible vindication that, yes, this absolutely works. And I’m excited to see it deliver even better results for our partners.