Test more than the ad

Every performance ad is a test. But not every test is useful. Not if all we measure is the click, the seven-day return, or whatever the platform's attribution model decides to take credit for this week. That's not a test. It's a carnival mirror, showing us a flattering distortion of what actually happened.

I've run plenty of ads that "worked": solid return, a nice bump in gifts, a click-through rate worth a high five. But a single-day spike tells you almost nothing about whether you grew.

What I actually watch

I track what happens beyond the ad. Email signups, time spent with the stories that matter, people coming back unprompted. I watch whether searches for the organization's name lift over time, because if nobody is looking you up more than before, the ad wasn't as sticky as the dashboard claims. And I follow donor cohorts: of the people an ad brought in, how many gave again at three months, at six, or became monthly givers. That number tells me more than any launch-week report.

My tracking is scrappy and partly manual, and I've made peace with that. The advantage isn't the tooling. It's knowing what you're looking for. If the goal is memory and trust, you can't stop at "return is up, therefore all is well."

The best test of this month's ad is whether next month's ad has an easier job.

There's good research behind this, too: the ads that build memory are the ones with character, humor, story, and human connection, the qualities that earn real attention instead of tolerating the skip button. That's where brand and performance stop being departments and become one discipline. So yes, test everything. Just test what lingers, not only what converts.