Purpose:
At ROI Solutions, our MiLo Intelligence team has reached new heights in exploring new and innovative ways machine learning can be leveraged for good based on real-world fundraising problems. The most recent exploration involves the development of models to support an organization’s mid-level giving programs. The model predicts a basic-level donor’s likelihood of upgrading to mid-level giving status. It prioritizes donors within the mid-level pool to help focus and improve your organization’s strategic development efforts on those most likely to make substantial mid-level gifts.
Usage:
Organizations with mid-level programs typically have more personalized and costly donor cultivation programs (different packages, special invites, etc.), and they try to populate those programs with donors with strong giving potential. An organization trying to expand its mid-level program often starts by populating prospect pools from basic membership files and using various outreach tactics to nurture these prospects over time. They typically populate these mid-level pools using basic filters, like recent giving amounts that approach mid-level. Many organizations would like to grow their mid-level programs faster but often lack the tools to effectively identify enough prospects. Our mid-level machine learning model can look at many more donor characteristics to predict who will most likely make a mid-level gift at some point in the future. From a modeling standpoint, what matters most is seeing donors move from basic-level to mid-level giving over time. The model assists in pinpointing the best candidates for mid-level cultivation by looking at campaign responsiveness and the aggregate benefit of longer-term advancement.
Our experience with modeling in this area has yielded solid results as measured in 3 different ways:
-
- Ranking mid-level donors based on their likelihood to be responsive to the campaign (who are the best people to receive this campaign).
-
- Ranking mid-level donors based on their longer-term aggregate giving (who are the best people to include in the mid-level program).
-
- Ranking basic donors based on their likelihood of giving a mid-level gift (who are the best people with no history of mid-level giving to consider including in mid-level campaigns).
When developing any new model, it is important to back-test against historical data to confirm that it meets its objectives.
As shown in the table, the model reveals that higher-value mid-level gifts are concentrated in the top classes, with over 90% of $1,000 gifts falling within the top class.
Effectiveness:
The mid-level model has proven effective in ranking donors based on their potential for subsequent giving and their response to specific campaigns. Subsequent giving from the 25,000 MiLo Class 1 donors who were part of the October 2018 Mid-Level Appeal Campaign significantly exceeded the initial campaign contributions. The model’s ability to rank donors based on their likelihood of making mid-level gifts and overall giving potential demonstrates its usefulness in identifying top candidates for your mid-level program.
Overall, the model is highly effective at pinpointing donors who will most likely give mid-level contributions of $500 or more. By accurately identifying these high-potential donors, the model provides valuable insights that help organizations focus on those most likely to make significant mid-level gifts. This capability makes the model an essential tool for enhancing and expanding mid-level giving programs.
MiLo Intelligence is designed to support your efforts in growing your mid-level giving program with maximum efficiency. By leveraging its advanced predictive capabilities, our model ensures that you can strategically target and nurture the most promising donors, ultimately optimizing your growth tactics and increasing the success of your mid-level fundraising initiatives.
Would you like to learn more about the work we’re doing at MiLo Intelligence? Our models are developed with particular client objectives in mind, so you might have the seeds for our next effective model based on your organization’s needs. Let’s talk!
