On a spreadsheet, every house-file mailing that clears “revenue > cost” looks like a win. Under scrutiny, a lot of that “win” evaporates.

In two large field experiments, simply telling donors to expect another campaign later reduced current giving.  The control condition raised about 60% more than letters that primed for future asks, and the later wave didn’t make up the loss. That’s called intertemporal crowding: you’re moving money across time, not creating it.

If you’re mailing your donors 12, 16, even 20+ times a year (as most large and mid-size charities do), do you think they’re surprised when another appeal shows up? Of course not. This experiment is your baseline, in living color.  The dollars showing up in your spreadsheet on the revenue line aren’t new, incremental ones, they are mostly shifted forward.

If this weren’t problem enough, you also face the stiff, unavoidable headwind of diminishing returns. Sector-wide panels show fundraising is productive on average, but the marginal dollar works less hard as you add volume.  Said differently, effort and cost to produce the next appeal are the same but returns are lower.

And it isn’t just inefficient. In a 60k person randomized experiment, one extra reminder raised gifts and sharply increased unsubscribes; the authors estimate an “annoyance cost” per reminder and show the net for the charity can be near zero or negative once future attrition is priced in.

“Always Be Asking” isn’t a timing strategy

Donor behavior isn’t random—it’s rhythmic. There are peaks when a person is more likely to act, and troughs when more asking just adds noise and often, those peaks align with temporal landmarks – fresh starts, holidays, anniversaries, birthdays, seasonal. Riding those landmarks beats spraying asks across the calendar.

We’ll be talking about a better approach in tomorrow’s webinar, this is your last chance to register,  Stop Asking, Start Aligning: How Donor Rhythms Beat Default Frequency

Here’s a preview of what we’ll discuss.

1) Split your file into two universes

Universe A: Annual-only givers (the “mode of 1” universe).  Hint: it’s a very big segment.
These donors have only ever given once per calendar year.

• Plan: One primary ask in their natural window (anniversary, fiscal year-end, or calendar year-end), two light set-ups, then gratitude/impact. Suppress hard asks outside that window.

• Why it works: You stop trying to manufacture second gifts that aren’t in their pattern, avoiding irritation and future-wave cannibalization documented in the mail-frequency and reminder studies.

Universe B: Multi-givers.
These donors have given >1× in at least one prior year.

• Plan: Build an individual-level timing model that uses both giving history and promotion history (what you sent, when, and how they responded). Concentrate hard asks in predicted windows; outside them, default to value or silence. After a gift, move to reinforcing their connection and competence, then re-enter the model.

• Why it works: Readiness clusters. Temporal models predict who is likely to give again and when; dynamic response modeling shows promotion history matters for next response. You’re matching cadence to personal hazard curves, not a broadcast calendar.

2) Change the scoreboard

For every wave, report:

• Net incremental revenue per 1,000 mailed (vs. a randomized holdout through the next scheduled campaign). This catches pull-forward.

• 12–24-month retention/LTV deltas for exposed vs. holdout. Sector benchmarks show short-term dollars can mask long-term erosion.

3) Rebuild the contact policy around evidence

• In-window (modeled peak): send your best hard ask for that person.

• Out-of-window: build their sense of competence and connection + go silent (give them a break)

• Post-gift: build sense of competeence and connection; suppress hard asks until their modeled next-likely window.

Kevin

P.S.