Before You Call Them Lapsed

The answer you can explain

Subtracting the 20 new donors from 75 gives 55. That includes ten returning donors who were never in the 2024 cohort, so the shortcut overstates retention by ten points.

A paragraph for the board report

The board needs the result in plain words. Whoever checks the figure needs the method. Keep them apart.

If you don’t have enough history to separate first-ever from returning donors, the board paragraph says “30 donors who did not give in 2024; we are still checking their earlier history.” Don’t label them all new.

When the mix changes, the rate moves

Overall retention can fall even when every group of donors behaves exactly as before. If first-year donors return at a lower rate than donors with a longer history, a prior cohort with more first-year donors produces a lower overall rate. This fictional arithmetic example keeps both group rates fixed:

The overall rate falls seven points, and neither group did anything differently. A year with more first-time donors can produce a cohort like B. Read alone, 46% looks like a stewardship problem. Read by group, it moves the board’s question from “what went wrong?” to “how do we keep more first-year donors?”, and it points effort at the second gift rather than at repeat-donor stewardship that is holding steady.

When the prior cohort’s mix has shifted, report the first-year and repeat rates, with their counts, beside the overall rate. First-year donors are prior-period donors with no earlier gift, so each split needs complete history before the prior period. The comparison tool below gives the split for one pair of periods when you mark that history complete; run it once for each pair you want to compare. The table above is worked arithmetic, not tool output.

Get two lists that mean the same thing

Before exporting, complete this note with the person who owns the report:

Use received gifts rather than unpaid pledges for this calculation. If you need a different metric, name it separately. For an unfinished year, use a clearly defined year-to-date comparison with the same cutoff in the previous cycle; do not present it as completed annual retention.

Export one stable donor ID per line for each period. Keep names, emails and gift notes out of the comparison lists. Duplicates in transaction exports should not create extra donors.

Know which history each label needs

Periods that aren’t adjacent leave a gap nobody has observed. Compare 2023 with 2025, and a current donor with no gift in 2023 or earlier may still have given in 2024. Check the gap year too, or report those donors as not yet classified.

Calendar boundaries create absences of their own. A donor who gave on December 20, 2024 and again on January 5, 2026 gave about a year apart, yet counts as absent from 2025. The rate for 2025 stays as it is, but that donor isn’t a candidate for lapsed-donor outreach.

Check the match with a tiny example

The intersection is EXAMPLE-102 and EXAMPLE-104: 2 ÷ 4 = 50%. EXAMPLE-101 and EXAMPLE-103 are absent from the current list. EXAMPLE-105 needs older history before you can call it new or returning.

If the prior-period list is empty, retention is undefined. Report “no eligible prior cohort,” not 0%.

The comparison matches exact, case-sensitive IDs, ignores duplicate lines and trims surrounding spaces. It won’t guess that two different IDs are the same donor. If a migration changed IDs, reconcile the old and current IDs in your records, then paste the corrected lists before relying on the result. Use these CRM migration acceptance checks to check record counts, attribution, and history before comparing periods.

Turn a missing ID into an investigation

A completed record for one absent donor:

Close the review when every absent ID has a finding and an owner. Then report the count for each finding next to the rate, so the board sees how many absences were payments, record changes or timing.