Growth A/B tests

Smart suggested amounts in the Donation Form Element

An updated version of our AI model grew recurring gift size but lowered conversion.

What we tested

Suggested amounts influence how much donors give. In the Donation Form Element, Fundraise Up uses AI to personalize these amounts in real time, adapting to donor behavior, device, and historical patterns.

To make sure we’re always showing the most effective asks, we regularly test new AI models against our existing version. This was one of seven suggested amounts experiments in 2024 across Checkout Modal, Campaign Pages, and Donation Form Element.

In this test, we focused specifically on recurring donation amounts in the Donation Form Element, aiming to see whether a new AI model would drive larger monthly or annual commitments.

Hypothesis

If the new AI model personalizes donation amounts more effectively, the average recurring donation amount will increase by 4% without negatively affecting the conversion rate, because the model will better tailor suggested amounts to donor characteristics.

Test setup

Duration65 days
TimingNovember-December 2024
Audience1.6 million visitors viewed the Donation Form Element
Group split50% control (current AI model), 50% test (new AI model)
Key metrics• Donation conversion rate • Average donation amount • Average revenue per user (ARPU)

Key results

  • −2.88% decrease in recurring conversion rate
  • +4.57% increase in average donation amount for recurring gifts
    • Notably strong performance on Android: +11.7%
  • ARPU remained flat overall
All key metrics passed the threshold for statistical significance (p < 0.05), confirming that these results are reliable.
2.88%
decrease in recurring conversion rate
4.57%
increase in average recurring donation amount
~
ARPU remained flat

Conclusion

The new model succeeded in increasing the average size of recurring gifts, but the drop in conversion offset those gains, leading to no meaningful lift in ARPU.

Still, these findings help us calibrate future model adjustments. In particular, we’ll explore how to balance gift size and conversion rate more effectively across device types and donation flows.

Even a small trade-off between fewer donors and larger gifts can have implications, and our job is to find the optimal middle ground.

What this means for you

Fundraise Up’s AI doesn’t stand still. Every day, we’re running experiments behind the scenes to find what truly works, so your donors are more likely to give, give more, and give again.

This experiment didn’t deliver a clear win, but it brought us one step closer to smarter personalization in every context. That’s the foundation of innovation and a better donor experience.

And while not every test leads to a rollout, the system as a whole keeps improving. That’s why Fundraise Up customers consistently outperform the sector:

  • +38% higher average recurring gift ($33 vs. $24 industry average)
  • 2–3x higher donation conversion rates
  • +25% better recurring donor retention (76% vs. 61%)

This level of performance is the result of continuous testing, learning, and optimizing.

→ Explore how to enable smart suggested amounts today

See how data-backed features drive better fundraising results

Request a demo