Non-ProfitFree
Donor Retention Analysis for Nonprofits
Send the CRM giving history, and get retention cut by gift sequence and acquisition source rather than as one blended annual rate.
Donor retention analysis usually produces one number. Last year's donors who gave again this year, divided by last year's donors, expressed as a percentage and set against a sector average. It is a real measurement and it is almost useless, because a single rate blends two completely different populations. Long-standing donors renew at one rate and first-time donors renew at another, and averaging them produces a figure that describes neither and points at nothing you could go and fix.
The Thornbury Trust is an invented example. It reports 48.79 percent retention: 5,450 of 11,170 donors gave again. Split the file and the number falls apart. Of 6,350 established donors, 4,318 renewed, which is 68.00 percent. Of 4,820 donors acquired that year, 1,132 gave a second gift, which is 23.49 percent. So 3,688 people gave once and never again, and they are 76.51 percent of the year's new donors. None of that is visible in 48.79 percent.
Then cut it by gift sequence rather than by year. Of the 1,132 who made a second gift, 852 made a third, which is 75.27 percent. The first-to-second step loses three-quarters of the cohort and the second-to-third step keeps three-quarters of it. That is not a retention problem spread across the file. It is one specific gate, and it is the only place worth spending on: five points on the first-to-second rate is 241 more donors, where five points on second-to-third is 57.
The cut that matters is by source, and the ranking flips
Sequence tells you where the leak is. Acquisition source tells you where it comes from, and the two answers disagree. Thornbury's peer-to-peer donors convert to a second gift at 10.00 percent, the worst rate on the file, but they are only 960 of the 4,820 acquired, so they account for 864 of the 3,688 losses, or 23.43 percent. Direct mail converts at 22.00 percent, twice as well, and accounts for 1,443 losses, or 39.13 percent. Worst rate and biggest hole are different sources.
Which is why the source cut has to carry both the rate and the count. Bringing peer-to-peer up to the 40.00 percent that web and organic already achieves would add 288 second-gift donors. Doing the same for direct mail would add 333. Meanwhile the 250 monthly sign-ups convert at 94.00 percent on a 22 dollar first gift, while 1,240 event attendees convert at 15.00 percent on a 185 dollar one. A first-gift-revenue ranking of those sources is the exact inverse of a retention ranking.
One data caution before any of it. Event gifts are usually part payment for something. The IRS requires a charity receiving a payment over 75 dollars partly as a contribution to provide the donor with a good faith estimate of the fair market value of the goods or services. If the gala ticket posts at its full face value, the event cohort's gift sizes are overstated. Recurring gifts distort the other way, because separate contributions of less than 250 dollars are recorded gift by gift.
How it works
Send the history
The gift-level export, several years if you have it, with source and appeal codes intact.
Build the cohorts
Donors grouped by acquisition year and source, then tracked forward gift by gift.
Find the gate
Conversion at each sequence step, so the leak is located rather than averaged across the file.
List the people
Lapsed and at-risk donors named, each judged against their own historical giving interval.
What you get
- Retention by gift sequence: first to second, second to third, and onward, with the gate named
- Every acquisition cohort cut by source, carrying both the conversion rate and the count of donors lost
- The whole-file rate decomposed into its established donor and newly acquired donor halves
- A lapsed list and an at-risk list, built from each donor's own giving interval
- Recurring and event gifts handled separately, so twelve monthly gifts are not twelve donors
- A doc that names the one lifecycle gate worth intervening at, with the arithmetic behind the choice
Common questions
Why is a single retention rate not enough?
Because it averages populations that behave nothing alike. Thornbury's 48.79 percent is 68.00 percent among established donors and 23.49 percent among first-time ones. The blended figure moves whenever acquisition volume moves, so a good acquisition year can make retention look worse while nothing about donor behaviour changed.
How do you decide who counts as lapsed?
From each donor's own pattern rather than a fixed thirteen or eighteen months. A donor who has given every November for nine years is not lapsed in June and is very lapsed the following January. A quarterly giver is at risk after five months. One global cutoff mislabels both.
Our CRM data is a mess. Does this still work?
It works better once the file is clean, and the analysis says which fields it could not trust rather than quietly using them. Missing source codes are the usual blocker for the cohort cut. CRM data cleanup is the step that comes first when acquisition source is blank on a large share of records.
Why do recurring donors need separate handling?
Because a monthly gift posts twelve rows a year and would otherwise read as a donor with twelve gifts and perfect retention. Thornbury's 250 monthly sign-ups convert at 94.00 percent, which is real, but it belongs in its own column. Mixing them into the file inflates every sequence step above it.
Does it tell us how much to spend on retention?
No. It sizes the gate and counts the donors behind it, which is what a budget conversation needs as input. Five points on the first-to-second step is 241 donors at Thornbury and five points on the second-to-third step is 57. What that is worth spending is a decision for the fundraising team and the board.
How does this connect to the appeal itself?
Retention analysis says which cohort to write to. The annual appeal and campaign pack is where the segmented ask gets written, and a second-gift ask to a first-time donor is a genuinely different letter from a renewal to a nine-year donor. The analysis is what makes that segmentation defensible.
Where do the gift records need to come from?
The CRM, reconciled. The gift reconciliation pack is what makes the giving history agree with the bank, and donor acknowledgment and receipting is where the deductible portion of an event gift gets separated from the ticket value that distorts this analysis.
Donor Retention Analysis for Nonprofits
Fill in the form and your workspace opens with the work already underway.