Monthly donors are extremely important supporters for any non-profit. Statistically, they give more over the course of a year than a one-time donor, and perhaps more importantly, they are usually a non-profit’s biggest champions through other forms of engagement. Can AI encourage these donors to maintain their monthly giving? Yes.
A new use-case is releasing to our clients in September for ‘monthly donor retention’. Call it like it is.
The models built for this use-case required a very different approach to problem solving; they are a significant departure from our previous work. Our team needed to rely primarily on the transactional history of each supporter to make a prediction on the likelihood they would lapse in the next 60 days. These ‘at risk’ donors were clustered and the content recommendations for the email automations were based on both transactional history and content preferences.
So what do the numbers show? Here is a breakdown:
- Let’s say a non-profit has 1,000 monthly donors
- We know from client data that this organization is likely to lose 7% of these monthly donors in the next 60 days – roughly 70 monthly donors; although this number varies of course between organizations
- Our model will identify about 240 monthly donors ‘at risk’ of lapsing over the next 60 days
- These 240 supporters will receive specific content encouraging them to maintain their monthly giving
- The client will have an opportunity to prevent about 35 from lapsing
Your first question is why only 35 and not all 70? Because the model will not identify 100% of the records that will ultimately lapse in the next 60 days. However, if the non-profit saves even 15 of these monthly donors, based on an average annual giving of $180, the non-profit will save almost $17,000 in annual revenue.
The accuracy of the prediction of ‘at risk’ donors will continue to improve as the models learn and more client data is used in the custom models. This is just the exciting launch baseline!
