The Fred Hollows Foundation NZ saved an estimated 211 regular givers from churning over a 7 month period, using Dataro’s AI prediction software and a well-timed intervention thank you calling campaign. With Dataro’s AI-powered donor churn propensity scores implemented directly into their CRM, The Foundation were able to identify which regular givers were most at risk of churning. Armed with this knowledge, the Engagement Team implemented a new monthly thank you calling campaign, targeting 150 at-risk donors each month (on average) and following them up with a personalised call to thank them for their ongoing support and re-engage them with The Foundation’s vision and purpose. Assuming each retained donor would go on to give an additional 18 monthly gifts, the intervention calling campaign will raise an estimated +$104k in net income in monthly gifts.

Key results

• Over 211 regular givers saved from churning in 7 months

• Raising an estimated $104,005 in net income through monthly gifts

• Estimated 28x return on investment
  
  

How we did it

Dataro’s Predict Regular Giving propensity model was implemented with The Fred Hollows Foundation NZ’s CRM – BlackBaud’s Raiser’s Edge NXT. Using machine learning, the Dataro model generated propensity scores for all of The Foundation’s active regular givers. These scores were seamlessly integrated in their CRM against the donor record, and showed how likely each donor was to miss three gifts in a row in the next 6 months or how likely they were to call and cancel their recurring gifts.

Using the list function, the Engagement Team were able to easily identify the 150 regular givers at risk of churn each month (on average) and provide a list of donors to their calling team for thank you calls.

Analysing the Results

Seven months after the calling campaign was implemented, we saw:  

1. Dataro’s predictions were accurate, with higher churn scores accurately correlating with a higher churn rate

2. Most regular givers appreciated the call with very few cancellations

3. Donors who received the call had a significantly lower churn rate compared to donors who did not receive the Thank You call

Critically, the results showed:

• Organisations not using machine learning are unable to identify which regular givers are at the greatest risk of cancelling their gifts. As a result, either no intervention tactics are deployed or the wrong people are included in generic thank you / impact reporting campaigns (increasing waste).
  
  

• For organisations who don’t have churn propensity scores, at-risk donors can unfortunately also be included in RG upgrade campaigns. This could potentially increase the likelihood of a donor cancelling their gift – the opposite result of what was intended!