How many of you have read performance reports where the email cost is $0?   Does anyone else find zero cost to be silly?

A study analyzed 746,426 email solicitations sent to 21,451 online survey panelists over three years.  The modeled out 3 possible explanations for attrition:

• people never permanently disengage
• people disengage as time passes
• solicitations themselves cause disengagement.

Guess which model best fit reality?   Yeah, repeat solicitations.  This is a finding that should surprise nobody save for those who believe $0 is the true cost or that mail more/make more is the correct growth model.

Each new email solicitation temporarily reduced the likelihood of responding to the next one. That effect faded with time but each email also increased the likelihood that the person had permanently checked out, even though they had not unsubscribed or announced their departure.

The underlying principle is simple: every solicitation can consume some of the donor relationship.

Another finding is the people most likely to respond were also more likely to disengage. The “obvious” targeting strategy of continually emailing the people who respond is a lousy long-term play.

In simulations, a “greedy” policy focused on previous responders generated 2,768 responses. Random targeting generated 2,941. An optimized policy that considered previous responses, solicitation volume and timing generated 3,845, a 30.7% improvement over random.

The supposedly smart strategy did worse than throwing darts.

The fundraising parallels abound.  Here’s my short list

• An appeal has a cost even when production and delivery cost almost nothing. Email is free in roughly the same way the hotel minibar is free.
• Campaign reporting captures the gift generated today but never captures the future gifts lost because another donor stopped paying attention.
• Unsubscribes do not solve the measurement problem because most attrition is silent.
• Recency, frequency and previous giving identify donors likely to respond but has zero to do with how much contact pressure those donors can tolerate. A model that predicts only the next gift can systematically exhaust the people it values most.
• Cadence needs to be treated as an individual decision, not a campaign calendar. The relevant question is whether this donor should receive this communication now, given that donor’s behavior in response to the number and timing of previous contacts.

A given cadence (8 mailings, 12, 24, 36) can “win” every appeal while damaging the donor file.

Kevin