Segmenting a donor list beats one letter for everyone

Aug 29
3 min read

Updated: 4 days ago

Introduction

The appeal goes to everybody. The person who gave once two years ago receives the same letter as the volunteer who has supported the organisation for a decade and gives monthly.

Both letters are wrong. One asks a stranger for commitment they have not built to, and the other asks a devoted supporter for a gift smaller than they would have given, while telling them things they already know. Segmentation is often presented as an advanced technique, and the useful version is simpler than that.

1. Segmenting a donor list can start with two groups

Do not wait for a sophisticated model.

New donors and established donors, treated differently, already outperforms sending everybody the same thing. The organisations that never segment are usually waiting for a level of capability they do not need, and lose several years of better results while they wait for it.

2. Split by relationship stage first

The division that matters most.

Prospective, first-time, repeat, lapsed, and regular givers each need a different message. A first-time donor needs reassurance and reporting back; a lapsed one needs re-engagement; a monthly giver should not receive an acquisition appeal at all.

3. Treat monthly donors as a separate group always

The most common and most damaging omission.

Sending a regular giver an appeal that ignores their existing support suggests nobody knows they give. At minimum, acknowledge the monthly gift and ask for something different, such as an additional one-off or an upgrade.

4. Vary the ask amount by giving history

The simplest change with the clearest effect.

Suggested amounts based roughly on what somebody has given before outperform a single set of figures for everybody. Someone who gave two hundred last year should not be looking at options starting at five.

5. Segment by how they arrived

Source predicts response more than most people expect.

Event attendees, emergency appeal donors, community fundraisers and long-term supporters behave differently and respond to different framing. Recording the source at the point of the first gift is what makes this possible later.

6. Respect stated preferences as a segment

Both a courtesy and an obligation.

What each person agreed to receive, in which channel, at what frequency. Preferences are not a soft signal to work around, and in most jurisdictions ignoring them creates a compliance problem as well as an annoyance.

7. Keep the number of segments manageable

Over-segmentation stalls the work.

Four or five groups that actually receive different messages beat twenty that all end up receiving the same one because nobody had time to write them. Add complexity only when the current segments are being used properly.

8. Change something real between segments

Segmentation without differentiation is just a filter.

The opening, the ask amount, the story, and whether there is an ask at all. If the only difference is the salutation, the sorting has cost effort and delivered nothing.

9. Measure by segment, not overall

The measurement is what makes it improve.

Response rate, average gift and second-gift rate for each group separately. Overall figures routinely hide the fact that one segment carried the entire campaign while another cost more to contact than it returned, and that is precisely the information you need before the next one.

Keep the underlying data clean, because segmentation depends entirely on it. Missing sources, unrecorded dates and duplicate records make every segment approximate, and approximate segments produce messages aimed at the wrong people.

Conclusion

Start with two segments rather than waiting for a sophisticated system, because anything beats sending everybody the same letter.

Divide first by relationship stage, always separate monthly donors so they are never sent an acquisition appeal, vary suggested amounts by giving history, use the source of the first gift as a predictor, treat stated preferences as a segment you must honour, keep the number of groups small enough that each genuinely gets different content, change something substantive between them rather than only the greeting, and measure response and second gifts by segment rather than in total.