In 2018, ROI Solutions embarked on an exciting journey to craft its very own machine-learning platform, known as MiLo™. Our primary objective was to pioneer innovative, tailor-made models capable of making a tangible difference across the entire constituent lifecycle within nonprofit organizations. Fueled by the same unwavering spirit of innovation and intellectual curiosity that has propelled our growth as a technology company, the MiLo team collaborates closely with our clients to pinpoint desired outcomes for any marketing or fundraising challenge.
With over two decades of expertise in nonprofit enterprise data management and direct marketing fundraising to guide us, ROI Solutions possesses a distinctive advantage when it comes to asking the right questions of MiLo, all informed by the insights and feedback of our valued clients. We enrich the platform with meticulously curated data and allow MiLo to perform its core function – identifying models that have a measurable impact on audience outcomes. MiLo boasts the capability to process thousands of data elements and interactions, deploying a range of well-established statistical models to uncover client-specific predictive data that will deliver optimal results for the organization.
Our unique approach allows us to do several things very efficiently:
- Collaborate with clients to understand the fundamental issue they are trying to solve in their fundraising and marketing programs, compare to what they have done in the past, and devise new best practices to engage with an optimized audience without a heavy lift for the organization.
- Allow MiLo to leverage each organization’s unique data to identify the best model to optimize audience selection for each particular use case.
- Continually fine-tune models via feedback from real-world results.
- Schedule and deliver scores to meet the marketing needs of our clients.
- Our familiarity with nonprofit data allows ROI Solutions to take on a lot of the tasks of model building and deploying freeing your team from this time-consuming work.
“The processing capabilities of the MiLo technology now allows organizations to finally leverage the richness and granularity of data that organizations have been collecting for many years, and help marketers get creative in activating it for positive outcomes for their organization in a really flexible way and rapid fashion.”
Lee Gartley
MiLo Team Lead
ROI Solutions
ROI Solutions partnered with clients to build, test, and refine models for various specific use cases. Unlike typical Coop-driven models, MiLo uses clients’ own unique data, both granular and summarized views, which allow MiLo to find the right combination and interaction of variables to predict and stratify an audience. Our familiarity with nonprofit data allows organizations to take advantage of machine learning modeling without having to spend a significant amount of time managing and curating data. Let MiLo do all the work for you!
The data leveraged goes far beyond the traditional cross-organization giving variables typically employed in a Coop model. MiLo considers not only transactional data, but also contact history, other interactions and engagement points, and even appended demographic data. Each model we deliver is client-specific because for each client the data that is found to be predictive is unique.
MiLo models have consistently delivered impressive results throughout the various iterations we’ve developed for our clients. At ROI Solutions, our commitment to innovation remains unwavering, and we are actively collaborating with additional organizations to uncover fresh inquiries that can be explored through our MiLo models, spanning the entire constituent lifecycle across a diverse range of communication channels.
Current MiLo Modeling Portfolio
Model | Description |
|---|---|
Appeal Modeling | Find the best responders in your active file to optimize costs while maintaining response rates. |
Lapsed and Deep Lapsed Recapture Model | Find pockets of responsive donors in lapsed & deep lapsed audiences. |
Mid-Level Upgrade | Identify donors to target for an upgrade into mid-level programs. |
Sustainer Loyalty | Identify sustainers at the highest risk of churn to improve retention rates. |
Sustainer Conversion | Identify constituents who will respond to sustainer asks. |
Targeted Ask Arrays | Maximize revenue per donor by asking for the right amount. |
Sustainer Loyalty Model
PETA Foundation
The Challenge
For organizations that have invested heavily in building their monthly sustainer programs, attrition can represent a significant revenue decline. PETA Foundation estimated that the attrition of their monthly donors could be as much as $1,000,000 annually.
The Model Build
The MiLo team worked with PETA Foundation to curate data for their 52,000 monthly sustainers. The Sustainer Loyalty Model identified data points of historical pledge performance to predict future pledge performance.
The Results
The MiLo Sustainer Model effectively predicted the most loyal sustainers and those most likely to churn. PETA Foundation is actively testing intervention strategies for the at-risk population and upgrade strategies for the most loyal.
Ongoing Strategy
PETA Foundation participates in ongoing cohort meetings with other organizations using MiLo’s Sustainer Loyalty Model to share best practices on intervention strategies to reduce overall churn.
Lapsed & Deep Lapsed Recapture
Human Rights Campaign
The Challenge
Human Rights Campaign (HRC) had been using a variety of Coop and internal driven models with varying results. At the start of the pandemic, they knew they had to try something new to recapture more lapsed donors to maximize fundraising efforts.
The Model Build
The MiLo team worked with HRC to understand typical audiences for existing lapsed recapture efforts and prior contact cadence, and then curated giving, response data and other engagement variables to create a robust model that was able to effectively extend HRC’s outreach to a far deeper lapsed audience.
+95% Revenue Lift
The HRC Lapsed and Deep Lapsed model uncovered some donors that the organization had not typically included in their lapsed recapture marketing programs. In fact, some of these donors had not been active in over 20 years!
The Results
The Lapsed and Deep Lapsed Model was tested against several internal and Coop models and performed extremely well. The response rate increased 41% while mail volume increased over 38% compared to prior efforts. The net result was a 95% revenue lift for HRC on their annual lapsed recapture program. HRC has adopted this model to replace their previous models and MiLo scores are refreshed prior to each campaign.
Lapsed Recapture Model
International Relief Client
The Challenge
An International Relief Organization wanted to expand their ability to reach lapsed donors while maintaining a stable response rate and average gift size. The overall goal was to drive incremental revenue without increasing cost per M.
The Model Build
The MiLo team developed a Lapsed Recapture Model in the lab using the client’s data with a focus on expanding the viable universe while stabilizing both response rate and average gift metrics.
The Results
Testing for the Lapsed Recapture Model at this organization has maintained overall response rate while simultaneously increasing the average gift size by about +6% against previous lapsed mailings. Using the MiLo Lapsed Recapture Model, the Client has been able to improve the targeting of their lapsed universe and shift focus away from straight acquisition to higher performing lapsed names. This has allowed them to raise over $200k more in donations while maintaining the same overall cost of the effort.
Increased Audience Reach
By expanding the universe of viable names at a set success metric of CPM, the organization was able to increase the overall size of their marketable lapsed pool by 50%
