Fundraising success relies not just on asking, but on asking the right donor, for the right amount, at the right time, through the right channel.

Sick Kids faced challenges optimizing its fundraising campaigns:

• When should they ask a donor for additional support?
• How much should they ask for to maximize giving without causing donor fatigue?
• Which communication channel would yield the best response for each donor?

Traditional segmentation techniques based on basic demographics and giving history were no longer sufficient to drive campaign performance at scale.

CVM worked closely with Sick Kids to develop a suite of AI-powered targeting models to optimize donor engagement across campaigns.

Key elements of the solution included:

• Ask Timing Models: Predicted the optimal window to solicit additional donations from individual donors.
• Ask Amount Models: Estimated the amount each donor was most likely to give in response to a solicitation.
• Channel Optimization Models: Recommended the most effective communication channel (e.g., email, direct mail, phone) for each donor based on historical preferences and engagement behaviors.
• Integrated Deployment: Delivered predictive scores directly into Sick Kids' marketing systems, enabling easy audience segmentation and personalized outreach.

The models were developed using Python, trained on historical donor interaction data, and built with flexibility to evolve as new data became available.

• Higher Campaign Response Rates: Improved personalization led to more effective fundraising campaigns across channels.
• Increased Donor Engagement: Donors received solicitations that were better timed, more appropriately sized, and delivered through their preferred channels.
• Optimized Marketing Spend: More efficient use of fundraising resources by focusing on high-propensity donors with tailored messaging.
• Scalable Targeting Framework: The modeling approach could be adapted for new campaigns, special appeals, and evolving donor behaviors.

• AI Enables Hyper-Personalization: Machine learning models can tailor outreach strategies to individual donors at scale.
• Actionability is Crucial: Predictive insights must be delivered in a way that marketing teams can easily use in real-world campaigns.
• Continuous Learning Drives Improvement: As donor behavior changes, regular retraining keeps targeting strategies effective.