We love tidy levers. Teens are anxious, so ban phones. Revenue is flat, so send more appeals. If the dashboard blips, we pat ourselves on the back. If it doesn’t, we reach for the same lever again.

Reality is not that simple.

Take the SMART Schools study in The Lancet.

• Thirty English secondary schools, 1,227 students.
• Restrictive schools barred recreational phone use; permissive schools allowed it at breaks.
• The outcome that matters, mental wellbeing, didn’t change – zero meaningful differences.
• What did change was the easiest to police: in-school phone time went down during school hours.  But total weekday and weekend use didn’t budge.

The system compensated outside the hallway we controlled, that’s the lesson. External forces (home routines, platform design, peer networks, sleep, parental boundaries) dwarf the narrow strip we can see. You can move the visible behavior at 11:17 a.m. and still not move the outcome at all.

Fundraising makes the same mistake every day.

Have you built a too-simple worldview?

Be honest: is your or your agency’s fundraising model basically this?

• “People give because we ask, so ask more in more places.” Volume stands in for strategy, if the metric twitches, we credit the extra touch. If it doesn’t, we add more touches.
• “Engagement” is what’s easy to count, not what it is. We chase opens, clicks, shares, comments, RSVPs.  Meanwhile, the most common by a country mile, high-intent behavior,reading, is barely measured. Attention is a better signal than clicks, and most “clickers” don’t read anyway.
• “Everyone’s the same.” One cadence, one message. Personality and identity differences are treated like abstraction while superficial demographics get high attention.

If your view is narrow, you’ve already shrunk your ability to influence outcomes. If you or your agency doesn’t understand the system, the parts you control and the parts you don’t, you misattribute causality to the lever you can see.

Control theater vs. causal change

Control theater feels great because it’s legible. You can show the memo, count the sends, and point to activity. But activity is not effect.

• Systems compensate. Kids reduced phone use in school, then shifted it elsewhere. Donors may reduce response in one channel if you flood another, or avoid you entirely when pressure rises. Field experiments show people pay a real cost just to avoid being asked. Over-solicitation can look like short-term lift while simultaneously baking in long-term loss.
• Outcomes sit downstream of “unseen” drivers. Brand trust, identity fit, pay-period timing, news salience, deliverability and inbox throttling, checkout friction. These forces are predictable even if they’re not on your dashboard.
• The base case is brutal. Retention remains low. If you can’t beat the counterfactual, what donors would have done without your extra touch, you’re shouting into the wind. In Q1 2025, the sector’s year-to-date retention was ~18%. You don’t fix that with a seventh reminder; you fix it by changing the experience.

What to do instead (the anti-hallway plan)

• Model the base case.
  Always hold out a true control: people who were eligible but did not receive the extra touch. If your incremental lift can’t beat natural giving, the touch is noise, not signal.
• Personalize by trait, not just transaction.
  Conscientious planners, empathic responders, novelty seekers—they need different stories, cadences, and ask logic. Psychologically matched messaging outperforms generic blasts in field settings reaching millions.
• Optimize journey, not volume.
  Decide who shouldn’t get the next ask yet to protect momentum and prevent churn. Over 6–12 months, this yields more total revenue than squeezing one more gift this week. The retention math demands it.

Kevin