How a Food Bank Partner Boosted Donor Conversion by 458% with Avid’s Propensity Model

There’s a lot that goes into deciding who should receive a specific appeal.

As a food bank, you want to fight appeal fatigue and reach the right people, so it’s logical to ask those who gave last time. They’re the people most likely to give to this campaign again.

But what if there are other people on your file you haven’t asked, and they would give, maybe even more generously? That’s where Avid’s propensity models can be leveraged.

This food bank, working with agency partner Brad Cecil & Associates, decided to test whether Avid could find the donors they were missing, and ran another test after seeing positive results across other food banks.

The Results

Avid’s Model Found 458% More Donors

  • 458% increase in conversion rate over the standard donor list
  • 24% conversion rate among donors Avid’s propensity model surfaced
  • 68% increase in average gift size among those donors
  • 99.2% level of confidence in the result

The campaign stayed the same.
The emails stayed the same.
The ask stayed the same.

The only variable was who received it, and that was enough to change the outcome entirely.

A propensity model uses data from all your sources—your CRM, donation platform, email engagement, and more—to create an audience of people most likely to take an action.

This is not just a “prediction.” It’s an analysis of behavioral patterns to find your next most likely donors.

The Challenge

The File Had More Donors Than the Campaign Was Reaching

Each year, this food bank runs a match campaign, where a matching gift multiplies every dollar donors give.

To protect their donors from appeal fatigue, they keep the audience tight. Instead of sending the appeal to their whole file, they send it only to people who gave to previous match campaigns. It’s a logical approach since those donors have already shown they respond to this kind of ask.

But a list built from past givers has a built-in blind spot.

Somewhere in the rest of the donor file are people who have never been asked to give to a match campaign, but who look a lot like the people who did. Some of them might give, and might even give more generously. The usual list never reaches them.

Before, finding donors who are more likely to give meant someone sitting down with donor history and giving patterns and building a model by hand, campaign by campaign. This food bank wanted a way to do this that didn’t require a data analyst and weeks of lead time.

The Solution

Let Avid Build the Audience, Then Run a Clean Test

1. One List In, a New Audience Out

Brad Cecil & Associates handed Avid a single input: everyone who’d given to the food bank’s last match campaign. Avid used that group as the training data for a propensity model, scanning the rest of the donor file for people who hadn’t given to that campaign but matched the behavioral profile of those who had.

2. Nothing Else Changed

To make sure the lift came from the audience and not the creative, both groups got the exact same campaign: same emails, same copy, same donation page. The propensity list and the standard list were the only variables in play.

3. The Difference was Immediate

The propensity-model converted  24% of donors. The standard list only converted 5%. That’s a 458% increase, with a 99.2% confidence level behind it.

The propensity model donors also gave larger gifts. The average gift size from the propensity-model group was 68% higher than the standard list. The model didn’t just find extra names. It found more generous donors sitting unreached in the file.

Under the old approach, those donors wouldn’t have gotten the email.

What’s Next

With two tests pointing the same direction, this food bank and Brad Cecil & Associates will use Avid’s propensity models in direct mail, where it can make an even greater impact by reducing the cost to acquire and activate a donor.

Use Avid’s Propensity Models for Your Next Campaign

Traditional modeled audiences are expensive. They often take a data scientist or a 3rd party contractor weeks to build, and they’re usually out of reach for anyone but the most well-resourced fundraising programs.

With Avid’s propensity models, that audience is ready in minutes. No extra fees, no contractor, no wait.

Build the list. Sync it to your channels. Launch the campaign.

Want to see how Avid’s propensity models work with your tools?
Schedule a tour of Avid to learn more »

In this article

A weekly read on fundraising, from inside the data.

Subscribe for Updates

A weekly read on fundraising.

Signal from inside the data—what’s shifting, what’s working, what we’re watching across 1,000+ nonprofits and $11B+ in annual giving. Delivered once a week.


Share Story