Early Stage Results

Our pilot performed better than expected. Early results (between 6 and 12 months) confirmed that donors with higher scores were more likely to migrate to mid- value giving. They had nearly double the migration rate compared to the control group; 5.55% (predictive model) vs 2.95% (control group) reached the HK$5,000 mid-value benchmark in 12 months. 

“Data and time are two of the most valuable assets for fundraising. A data-driven analytics tool with predictive power helps determine which donors to focus on for a faster and more cost-efficient conversion.”

This is a promising result as it shows clear potential for increasing mid-value giving through smarter targeting. However, we had a few challenges with time and project management, and the small donor database made it difficult to reliably forecast volumes, limiting the ability to scale confidently. 

The pilot was a key step forward for us as it combined data science with a robust A/B testing framework which enabled us to validate performance in a measurable and credible way. In the long term, it marked a shift towards more evidence based fundraising, helping us identify and prioritise the donors that have the highest potential to become mid-value givers. The model itself not only confirmed its predictive accuracy but also gave us practical insights into donor behaviour, timing and channel effectiveness. As we continue to grow our donor base and secure more resources, we aim to expand our approach using this data to guide smarter decisions and improve long term donor value.

What we would do differently 

• Consider waiting until the donor base was larger before launching a full-scale predictive model.
• Extending the measurement window beyond 12 months, as many donors reached the mid-value thresholds just after the 12-month mark, suggesting a longer runway was needed to fully capture donor migration.
• Diversifying outreach channels beyond direct mail such as using digital or phone touchpoints.

Advice for Reimagining Fundraising Members 

1. Invest in building the right testing framework from the start. Predictive modelling on its own is not new, but what made this pilot valuable was the ability to validate results through a clear A/B testing approach. It gave us confidence in what was working and where we needed to adjust.
2. Invest in building the right testing framework from the start. Also, be realistic about timelines. Mid-value donor migration often takes longer than expected, and a twelve-month window may not capture the full picture. Set expectations accordingly and allow space to learn over time.
3. Treat this as a  long-term investment in building capacity. Even if the immediate return is modest, the insights gained will strengthen your overall fundraising strategy.