Artificial Intelligence Will Make Your Nonprofit Smarter

As a recent blog in NonProfit PRO described, artificial intelligence (AI) is all the rage in the nonprofit world. Traditionally slow to embrace technologies employed in the for-profit world, nonprofits have actually been early adopters of AI.

Unlike some technologies, which burst onto the scene but then fail to live up to the hype, we believe that AI is here to stay and that it will become an increasingly important tool for fundraisers. AI points us toward more likely situations and prospects saving time, money and heartache. That’s why we have formalized a partnership with boodleAI, a fast-rising star in the application of AI to fundraising. Turnkey looks for, “what’s next?” This is what’s next.

AI is the next logical step in how we find and interact with supporters. It allows us to take the gigabytes of data that we have on current and potential constituents and makes it actionable instead of just making for good theoretical discussion in a boardroom.

Until now, nonprofits typically relied on demographic data to understand its supporters, creating the familiar “donor personas” based on information in their database. The for-profit world also utilizes customer personas extensively. For example, Facebook offers to “clone” a customer list based on current customer demographics to target advertising to a new group.

But targeting based on demographics has serious limitations that are rooted in psychology, limitations that goes to the core of why individuals behave in the ways that they do.

Predictability Ceiling

Most people believe that individual personality traits, what psychologists call “dispositions,” are responsible for the majority of a person’s behavior. Let’s say that Person A gives to a charity supporting people experiencing homelessness. We might infer that they do so because they are a generous person, someone who has empathy for the plight of those without homes. Or we decide that maybe they had been in that situation themselves sometime in their past. We like to believe that their behavior is rooted in the kind of person they are.

We might try to use this information to predict how Person A would behave in new situations, like donating to a health-related or an animal welfare organization. But it turns out that the type of information about personality that most laypeople would want to have before making a prediction would prove to be of relatively little value. A half century of research has taught psychologists that in novel situations, one cannot predict with any accuracy how a particular person will respond using information about an individual’s personal dispositions, or even about that individual’s past behavior.

This so-called “predictability ceiling” is typically about 30%. As an example, let’s say we have a personality test that reliably measures how Person A feels about cheating. Using this information, we could predict how often Person A would cheat (or not cheat) on a test only about 30% of the time. Moreover, the 30% value is an upper limit. For most novel behaviors in most domains, (think, responding to a request to donate to a nonprofit) psychologists cannot come close to that.

Despite such evidence, however, most people staunchly believe that individual differences or traits can be used to predict how people will behave in new situations. Such “dispositionism” is widespread in our culture. That’s why predicting who is likely to donate to a charity based on matching donor personas makes intuitive sense. Yes, it’s useful information, but placing too much faith in demographics to predict behavior is a fool’s errand.

Where AI Comes In

Here’s where AI steps into the picture. In psychologist-speak, if dispositions (personality traits) account for no more than 30% of a person’s response to a novel situation, what is responsible for the other 70%? The truth is that most behavior depends on the situation, the environment. Using a nonprofit’s database, AI is a tool that allows the nonprofit to deliver a message tailored to an individual in a way and at a time that makes it most likely that they will respond positively.

As an example, Turnkey analyzed data from 160,000 participants in 2017 peer-to-peer walk campaigns to determine who was most likely to activate to fundraise. Among what we found was that participants who provided an email address were 2.8 times more likely to activate, those who participated in the previous year were twice as likely to activate and those raising money for a large organization were twice as likely to activate compared to those registering for a walk benefitting a medium-sized organization. These data points are about the situation.

All pretty interesting stuff! The problem is, how do you make it actionable? AI tools use this type of data, in conjunction with demographic data, to tell the fundraising organization which individuals in their database to contact, when to do so and suggests what to say to them to maximize the chance they will respond positively. Using both kinds of data, the predictability ceiling is much higher.

That’s just one example of how AI is impacting the nonprofit sector. The bottom line is that AI will make nonprofits smarter and will provide them with an edge to fulfill on their missions. That’s why is AI in your future. Who wouldn’t want a smarter organization?