Reviews
Part of Donor acquisition: judging a channel by retained donors, consent and capacity
Setting a donor acquisition threshold: five steps and an example
Set a donor acquisition threshold your board can check: count net donors, compare retention like with like, and write the stop rule before the campaign.
What to take away
- A sector range shows where you sit among peers. Your own threshold decides whether to keep spending.
- Cost per acquired donor is spend divided by net new donors, with refunds and duplicates taken out.
- Compare retention with your own last comparable campaign, not a sector average with a different maturity window.
- Count only donors you may lawfully contact, and name the authority in your file.
- Set the threshold, the review date and the stop rule before the campaign, not after.
Five steps from a sector range to a threshold
Start with a sector range for context, then finish with a threshold built from your own file. The steps below keep the two apart, so a peer study never decides a spend that you can check against your own numbers.
- Define the populationwho counts as a new donor and who is excluded.
- Fix the period and the maturity window before a donor counts as retained.
- Work out net new donors and cost per acquired donor.
- Work out retention for the same cohort in a later window you name.
- Set the threshold, the review date and the stop rule, then write them in the campaign file.
Population decides most of the argument. New donors, first-time gifts, one-time gifts converted to monthly and all gifts above a floor give three different counts, so name the one you use.
Net new donors and cost per donor
Cost per acquired donor is spend divided by net new donors, where net excludes refunds and duplicates. Gross gifts flatter the channel that spends most, so count only the donors left after those are removed, and keep the same maturity window each time.
Say a spring campaign spent $12,000 and produced 400 net new donors after refunds and duplicates. Cost per donor is $12,000 divided by 400, or $30. What that means depends on the threshold your team agreed before the results arrived.
| Agreed threshold | Cost per donor | Decision |
|---|---|---|
| $35 | $30 | The channel continues |
| $25 | $30 | Pause the channel and look at the gift array instead |
The same $30 gives two answers because the threshold, not the number, carries the decision. That is why the threshold is written down first. Set it after the results arrive and it becomes a way of excusing whatever happened.
Retention: compare like with like
Retention is the share of the cohort still giving in a later window you name. If 120 of those 400 gave again in twelve months, retention is 30 percent. Compare it with the same figure from your last spring campaign, not a sector average.
Read cost and retention together. A channel that meets the cost threshold but loses its cohort quickly has not acquired donors in any useful sense. Donor retention work starts from the same cohort definition, so reuse the one in your file.
Where a sector range fits
Sector studies from AFP and M+R give ranges, not targets. A range tells you whether your cost per donor sits near peers. It does not tell you whether to keep spending, because your mission, your list and your gift array are not the study's.
External vs internal benchmark
External
- Source
- Sector study
- What it gives
- Range
- Decides action
- No
- When disagree
- Context only
Internal
- Source
- Your program data
- What it gives
- Threshold
- Decides action
- Yes
- When disagree
- Internal wins
Keep the two in separate columns. The external column cites the study, its year and its population. The internal column cites your own program data and the threshold your team agreed on. When they disagree, the internal column decides.
Recalculate the external figure where the study publishes its method. If it does not, treat the number as context and say so in the file. A benchmark with a different denominator is a different benchmark, so say so or drop it.
Count only donors you may contact
For Canadian organizations, a donor you cannot lawfully contact is hard to count as retained. Under the anti-spam statute, a person who alleges consent has the onus of proving it, so the burden sits with the sender and the record has to exist.
The CASL consent rules cover consent and implied consent windows, while PIPEDA and BC PIPA govern personal information handling, and Quebec adds Law 25 and Bill 96 requirements. Under PIPEDA, individuals can withdraw consent at any time, subject to legal or contractual restrictions and reasonable notice.
Name the authority in your file and confirm the current rule with a qualified charity lawyer or the CRA. Consent records, suppression lists and retention limits belong in the system, not in a donor's memory of what they ticked.
Write four fields into the campaign file
Four fields turn a number into a benchmark: population, period, measure and decision. Write them into the campaign file before you name any figure, so a board member reading it a year later can rebuild the number without asking you.
Four fields to record
- Populationwho is counted
- Periodstart, end, maturity window
- Measurecost per donor, retention
- Decisionthreshold set before results
Four fields written down
| Field | What to record | Test before you use it |
|---|---|---|
| Population | Eligibility, geography, segment, exclusions | Same people counted both times |
| Period | Start, end, maturity window, season | Same window length |
| Measure | Net new donors, later retention, cost | Same denominator and currency |
| Decision | Threshold, uncertainty, review date | Threshold set before results arrive |
The same four fields belong in your marketing analytics reporting, so each figure carries its definition into every report that quotes it.
Common questions
When should the threshold be set?
Before the results arrive, together with the review date and the stop rule. A threshold written after the campaign can only describe what already happened, so it cannot decide anything.
What if a sector number uses a different denominator?
Then it is a different benchmark. Say so in the file, or drop it. Comparing your cost per net new donor with a figure built on gross gifts flatters one side and misleads the board.
How do we handle a benchmark we cannot recalculate?
Record the source, its year and its population, then label it context rather than a target. Say in the file which figure you could not verify and why.
