.jpg)
One thing we know about our donors is their giving behavior. And as a result, fundraisers have traditionally used that data to make communication decisions. A donor gives through the mail, so we keep mailing them. A donor gives online, so we send them more email.
But after years of testing mail frequency with online donors, I think there is a more important question: Is the current transaction channel actually the right way to determine how a donor wants to engage with us?
Our testing suggests it isn't that simple.
We know direct mail can influence donors who ultimately give online. Our past testing has shown that reducing or eliminating mail can hurt retention and donor value. But we've also found that some online donors can receive significantly less mail without sacrificing net revenue.
Over the past year, we pushed that question further to optimize net revenue even more. If a donor consistently chooses to give online, could we also reduce the cost of the mail we do send by removing components like reply envelopes and bouncebacks?
And topline results told us we could.
But when we looked more closely at which donors were affected by receiving less mail, we found that just because someone gave their donations online does not mean they were telling us they wanted to change their communication preference with the organization.
The Test
For our latest test, we selected active donors who had given online and had not made a direct mail gift in the 24 months prior to their most recent gift.
We split them into three groups:
- Control verification panel
- Test 1: fewer direct mail appeals and reduced mail components
- Test 2: fewer direct mail appeals with the existing package
Donors in the test panels were eligible to receive six fewer direct mail appeals during the year.
We also tested removing package elements such as reply envelopes and bouncebacks. If these donors weren't actively giving through the mail, did every mailing need all the components designed to generate a mail response?
The mail reduction and design changes resulted in significant savings. Cultivation expenses in the two test groups were 27% to 32% lower than the control.
The panel receiving fewer appeals with reduced package components beat the control on net revenue. The panel receiving fewer appeals with the existing package essentially tied the control on net.
We spent considerably fewer dollars without sacrificing net revenue. But as we had seen in our previous testing, retention moved in the other direction.
Control retention was 24.5%, compared with roughly 23% in both reduced-mail groups.
At first, that seemed to reinforce what we'd seen before: even when donors give online, reducing mail can hurt retention. But when we looked underneath the topline result, something much more interesting emerged.
The retention decline wasn't consistent across all of these "online donors."
Among donors who had never made an offline gift, reducing the amount of mail sent had almost no impact on retention. Control retention was 14.8%, compared with 14.7% and 14.6% in the reduced-mail panels.
The larger differences appeared among donors who were giving online today but had a history of responding through the mail.
In other words, two donors could look exactly the same based on their recent transactions—they both give online—but have very different relationships with the organization's communication.
One may have been acquired online and may have never sent a gift through the mail. Another may have spent years responding to direct mail before eventually choosing to complete their gifts online.
They may both be "online donors." But they aren't necessarily the same kind of donor.
A donor can change how they give without changing through what channel, or how often, they want to hear from you. But that does not mean we should just keep sending mail when it's not delivering a positive net return to the audience.
What If We Reinvest the Savings?
Across these test panels alone, reducing mail generated roughly $90,000 in cultivation savings. On one hand, we could simply take the savings and accept the lower retention rates. Another option is to keep sending all the mail to protect retention even at a lower net.
But neither of those answers helps us build a healthier donor experience.
The real question is whether we can use those freed-up dollars to create a donor experience that beats the control. What happens if some of that $90,000 funds a more relational communication experience?
The reinvested dollars could create more opportunities for donors to tell us why they care about the work. Or be used to occasionally send handwritten communication without an ask attached. Or help fund a mid-level representative or a "caseload-lite" program. We tested a version of this for donors who aren't yet ready for traditional major-gift cultivation, and we saw retention increase.
From Channel Optimization to Donor Experience
The lesson from this test isn't that organizations should mail online donors less often to boost their net. Or that removing reply envelopes will save you some money. That's far too simplistic.
What we actually learned is that we were able to remove a meaningful amount of communication expense while preserving net revenue—and that gives us resources to experiment with building something different.
For years, fundraising programs have been organized around individual channels: direct mail, email, digital, radio, events. But donors don't experience our organizations that way.
They experience one relationship.
A donor's decision to complete a transaction online tells us how they chose to give that day. It doesn't necessarily tell us how they want to hear from us, what will make them feel connected to the mission, or what experience will make them more likely to stay.
That's what I want the next generation of testing to answer. Not simply:
How much direct mail can we remove?
But:
What combination of experiences creates a stronger donor relationship—and generates more resources for the mission?
That's a much bigger question than direct mail optimization. And it's where fundraising needs to go next.
