Engaging the Right Mid-Value Prospects and Donors to Support Displaced People Worldwide
Australia for UNHCR is the UN Refugee Agency’s national partner in Australia. It raises funds and awareness to support displaced people worldwide.
Since late 2021, the charity has used Dataro’s AI donor predictions to improve fundraising ROI in its appeals program.
In 2023, the UNHCR team wanted to see if they could use Dataro’s AI predictions to refine the focus of their growing mid-value program. Specifically, they wanted to identify the most valuable donors from their existing middle program and find new prospects from other areas of their supporter base.
In managing an already successful mid-value program of more than 8,000 donors, effectively prioritizing stewardship also presented a formidable challenge and opportunity to leverage AI donor insights.
The UNHCR team sought to leverage Dataro’s AI insights to:
- streamline their stewardship efforts and resources
- prioritize mid-value donors most likely to yield positive fundraising outcomes
Following 14 months of testing, the charity restructured its mid-value program creating new custom mid-value streams with personalized stewardship journeys, based on their propensity:
- high stream mid-value (premium, high-touch journey)
- standard stream mid-value (reduced cost, semi-automated journey)
- core mid-value (no stewardship for lapsed & low-propensity donors)
This new approach enabled more efficient resource allocation, allowing key mid-value donors to be prioritized while avoiding excessive focus on low-propensity mid-value donors who contribute less.
As a result, the charity has been able to scale the right-sized mid-value program for its needs.
Key Results
Using Dataro’s AI donor predictions, the UNHCR fundraising team were able to:
- reduce its middle program from +8,000 to +4,200 donors (receiving high-touch stewardship)
- reduce cost and resources for low propensity mid-value donor stewardship, without risking revenue
- improve donor retention and maintain net revenue by treating low propensity ‘lapsed’ middle donors as core (via lower asks and pack costs)
- enhance overall mid-value program effectiveness and save 3 hours per week in stewardship resource
- identify 327 new mid-value prospects in the existing database
- convert 20% of newly identified prospects to mid-value (irrespective of high or low propensity)
- lift conversion rate from 3% to 28% for high-propensity prospects from core supporter base
Background
Australia for UNHCR has an established mid-level program of 8,000+ donors who give between $1,000 and $10,000 annually.
The charity’s mid-value donors typically receive quarterly appeals (including telemarketing calls), emergency appeals and recurring stewardship communications such as newsletters, special events etc. They are also included in rolling recurring giving telemarketing campaigns.
Before using Dataro’s AI donor predictions, mid-value donors were classified into three broad categories:
- Top Tier (Active v Lapsed +24 Months)
- Low Tier (Active v Lapsed +24 Months)
- Prospects
Note: both active and lapsed donors were treated the same within tiers
Using Dataro’s propensity scores, the charity aimed to distinguish high-propensity from low-propensity donors (regardless of recency or tier) and tailor their mid-value engagement efforts accordingly.
The objective was to identify donors for prioritized high-touch stewardship, manage the workload more effectively, and maximize program ROI.
How we did it
In this case, Dataro’s Mid-Level donor propensity model was implemented by connecting to the charity’s Data Warehouse through SFTP file transfer.
Using Machine Learning, Dataro’s software generated propensities for all of the charity’s active cash donors, providing a prediction of how likely each donor would be to give between $1,000 and $10,000 cumulatively over the next 12 months if stewarded effectively. These are expressed as a rank (rank 1 is highest propensity, rank 10,000 is low propensity). Predictions are based on the potential value of gifts, not ‘who is going to give a cash gift (or not)’.
Test 1 – Low Propensity ‘Active’ Donors
For active mid-value donors with a low Dataro propensity, the charity wanted to understand the optimal level of stewardship required to maximize return on investment (ROI) while maintaining donor engagement.
Specifically, they sought to determine whether reducing stewardship efforts would lead to better ROI compared to increasing or maintaining the standard level of stewardship engagement.
These donors were divided into three streams:
- Reduced cost stewardship (incl lower cost packs, removed from TM)
- Increased stewardship (treated as high net worth with high-cost packs & calls)
- Control stewardship (standard mid-value journey)
Key findings:
- Reduced stewardship stream had the best ROI for low-propensity donors
- Increased stewardship is not viable (a small lift in income does not justify increased costs)
- Control had a higher net income per donor but required a significant increase in time and resources to achieve the extra $35
- Reducing the level of stewardship for low propensity mid-value donors can save 3 hours a week in staff cost
Analysis of Results:
The reduced stewardship stream yielded the best ROI, with only a slight decrease in net revenue per donor ($35 lower than control).
The results showed a significant 14-point higher ROI compared to the control group, indicating a substantial difference in gross income relative to resource costs. This highlights the effectiveness of prioritization and return on investment.
Donors in the increased stewardship stream didn’t give more than the control or reduced stewardship group to justify the increased level of stewardship (net revenue per donor was $200 lower than the control).
While the control group showed slightly higher income per donor, it required substantially more time and resources to achieve only an additional $35 per donor.
Increasing stewardship resulted in more frequent but lower-value gifts, and the added cost of stewardship (as measured by actions in Raiser’s’ Edge) did not translate to a significant increase in income or ROI.
Additional Testings:
Australia for UNHCR team also wanted to understand if Dataro’s mid-value propensities accurately identified donors as high propensity or low propensity and if high propensity mid-value donors went on to contribute more (regardless of active or lapsed status).
Can UNHCR identify low-propensity mid-value donors without Dataro?
The UNHCR team analyzed the distribution of low-propensity donors across their active and lapsed mid-value segments and found:
- A significant volume of low-propensity donors were found in active segments
- Donation frequency was comparative between low and high-propensity donors across channels
- Identifying low-propensity donors would not be possible without Dataro’s Machine Learning model
How accurate are Dataro’s mid-value donor predictions?
The charity also compared the outcomes for low-propensity donors (in the control group) to high-propensity donors (not in the testing campaign). Both received the same standard mid-value stewardship journey. The charity found:
- Dataro’s AI insights accurately predicted that high-propensity mid-value donors contribute more
- The standard mid-value journey returned strong ROI for high propensity ‘active’ donors (maintain strategy)
Key implications for mid-value strategy
Without Dataro’s AI insights, the charity wouldn’t have known whether to prioritize or deprioritize thousands of donors in their mid-value program. As a result they:
- Implemented reduced-cost mid-value stewardship journeys for most low-propensity ‘active’ donors
- Used additional attribute tagging to determine which low-propensity donors to keep in standard mid-value stewardship journey
- Incorporated Dataro’s insights into Power BI reporting to enable stewardship prioritization without integrating into the CRM
Test 2 – Low Propensity ‘Lapsed’ Donors
The charity wanted to understand the most cost-effective approach to re-engaging lapsed mid-value donors (+24 months) with a low Dataro propensity.
They tested moving low-propensity lapsed donors back to a core donor level of engagement (and asking for lower gift amounts to retain net revenue) versus keeping them in the more costly mid-level program.
This test aimed to optimize the allocation of resources (communication and staff) while maximizing revenue and boosting donor retention overall.
Key findings:
By moving low-propensity mid-value donors back to core donor levels of engagement and asking less, the charity maintained revenue while reducing costs and improving team efficiency.
They found that spending more (in time and cost) on lapsed low-propensity mid-value donors did not result in higher income.
As a result of testing Dataro’s AI predictions, the charity implemented the following:
- Treat low-propensity lapsed mid-value donors as core
- Treat high-propensity lapsed mid-value donors as mid (high stream mid-value)
Overall, this test reinforced the accuracy of Dataro’s scores and helped the charity optimize its resources.
Test 3 – High Propensity Mid-Value Prospects
Australia for UNHCR wanted to see if Dataro’s AI donor predictions could also help them identify new mid-value prospects to convert to mid-value via their standard mid-value conversion journey.
Using Dataro’s AI predictions, the charity identified 376 core donors with a high propensity to upgrade to the mid-value giving range. These donors received the standard mid-value stewardship journey for 12 months.
Key findings:
- 66 donors converted to mid-value after 12 months
- 28% conversion rate for high-propensity core donors (61 donors)
- 833% lift in mid-value conversion rate for core donors (usually 3%)
As a result of testing Dataro’s AI propensities, the charity decided to treat high-propensity mid-value prospects from core supporter base as mid-value for 12 months. Donors who gave at the mid-value level after 12 months moved into the mid-value segment of donors. If they lapsed after 24 months, they dropped back into the core segment.
The trial proved that Dataro’s donor predictions correctly identify donors most likely to move into the mid-value level of giving if sent on the standard mid-value stewardship conversion journey.
In Summary
Based on these results, Australia for UNHCR used Dataro’s AI donor predictions to segment their mid-value program of now +9,000 donors into three distinct mid-value streams with distinct donor journeys:
- High Stream Mid-Value (premium journey), +4,200
- high propensity middle or core donors + additional bespoke tagged donors
- receive original standard mid-value journey (higher touch, more premium packs)
- Standard Stream Mid-Value (reduced cost journey), +4,000
- low propensity ‘active’ middle donors
- receive semi-automated stewardship with lower cost mid-value packs
- Core Donor, +1,200
- low propensity ‘lapsed’ middle donors
- treated as core donors again for purposes of appeals (packs and ask amounts)
This strategy has enabled the charity to reduce its traditional mid-value program from +9,000 donors to +4,200 and ensure its high-touch stewardship efforts are aimed at the right donors.
The reduced cost journey for low propensity mid-value donors has enabled the charity to effectively engage these donors without risking revenue or expending excessive effort and resources on donors who are less likely to continue giving at a middle level.
This targeted approach has allowed the charity to maximize the impact of its mid-value program while optimizing resource allocation for more sustainable growth.
What they said
“With finite resources, and nearly 9,000 mid-value donors to manage, Dataro’s AI insights have allowed our stewardship team to identify and focus their efforts on the best donors, who are most likely to continue giving at a middle level. This data-led approach has allowed us to be more strategic in our prioritization and stewardship of mid-value donors which is ultimately leading to more efficient use of resources and more effective donor engagement. Without Dataro's AI donor predictions, we would have no way to identify and segment high-propensity donors from low-propensity donors, within our existing RFV segmentation model. Dataro’s AI insights allow us to deliver an optimized stewardship journey for the organization and the donor.”
Jo Purcell-Jones
Business Intelligence Manager
More Case Studies
Using AI to Raise More Funds that Improve the Health and Happiness of Bristol & Weston Patients
Bristol & Weston Hospital Charity is the NHS charity committed to helping their group of hospitals go the extra mile for patients and their families. Despite being a relatively small nonprofit, they are at the forefront of embracing cutting-edge technologies like AI in their fundraising efforts to help
Read moreChild Cancer Foundation Leverages AI Donor Insights to Reconnect with Lapsed Donors
The Child Cancer Foundation (CCF) has long been a lifeline for Kiwi children and families affected by cancer. Over the years, the organization has relied on its generous donors to fund its work supporting families with children on their cancer journey. In 2021, staff changes left CCF with
Read moreHow Save The Children used AI to engage more mid-value donors to support children in need
Save the Children is the world’s leading aid and development agency dedicated to protecting children and helping them reach their full potential. In 2022, Save The Children Netherlands partnered with Dataro to see if they could convert more of their mass-market donors (single giving, recurring giving, multi-donors) to
Read more$831k in Extra Funding Secured in One Year with AI Donor Predictions
A leading international nonprofit dedicated to improving the lives of people with disabilities in developing countries partnered with Dataro seeking better direct mail appeal returns. This charity relies on the success of its direct mail appeal program to fund its important work and implements approximately 12 campaigns per
Read more