They Know the River

Every autumn, along the rivers of the Pacific Northwest, something ancient and precise begins.

Salmon that have spent years at sea find their way back — through open ocean, through coastal waters, into the mouths of rivers they haven’t seen since they were fingerlings — and begin the long push upstream. They navigate by smell, by memory, by something encoded so deeply in their biology that researchers are still working out the full account of it. They know the river the way you know a thing you didn’t learn so much as become.

And when something blocks the passage — a dam, a logjam, a section of channel that has shifted since the last run — they try again. And again. The effort is not confused or frantic. It is methodical, exhausting, and completely faithful to the only instruction the salmon was ever given: get upstream.

What the salmon will not do — cannot do — is stop and ask why the passage is blocked.

That question requires a different kind of learning than the salmon was given.

Marcus had been a director of development for nine years when he sat down across from me at a regional AFP session last fall.

He was the kind of fundraiser other fundraisers described as “relational” — a word that, in the profession, carries real weight. He had donors who had followed him across two jobs. He had a major gift in his portfolio that had grown from a $5,000 annual gift to a six-figure pledge over the course of eight years of what he called “just paying attention.” He was good at this work. The results said so.

The session asked him to name a failed receipt — a promise his organization had made to a donor that hadn’t been kept. He didn’t have to think long.

“We lost a donor last year,” he said. “Seven years with us. Gave every year. And then she didn’t. We called. She was pleasant. She said she just felt like we’d moved on from her.”

He paused.

“We sent her a card. We upgraded her to a mid-level acknowledgment package. We made sure she was on the impact report list. She gave again the following year. Smaller. And then not at all.”

The session facilitator waited a moment, then asked: “Can you trace it?”

“Trace what?”

“The gap. Between what she experienced and what your organization intended.”

Marcus sat with that for a second. “We sent the card late. The card was supposed to go out within a week of her gift. It went out in three weeks. My assistant was covering two positions that month. We fixed it.”

“You sent the next card faster,” the facilitator said.

“Yes.”

“And the gap that produced the late card — the fact that one person was covering two positions, with no process for flagging when acknowledgment timelines were at risk — did that change?”

Marcus looked at the table for a moment.

“No,” he said. “I don’t think it did.”

He had fixed the card. He had not fixed the system that produced the late card. And the system, unexamined, went right on producing.

This is not a story about Marcus failing. Marcus is a skilled, attentive, genuinely relational fundraiser. He did exactly what his training taught him to do: he detected the error, he corrected the behavior, he repaired the relationship. The loop closed.

The problem is that the loop closed at the wrong level.

Chris Argyris, the Harvard Business School professor who spent decades studying how organizations learn, called this single-loop learning. You detect an error. You correct it. You move on. The behavior changes; the underlying system that produced the behavior goes unexamined. It is efficient, familiar, and — for stable, well-understood processes — entirely adequate.

The late acknowledgment went out. Marcus sent the next one faster. One loop.

What single-loop learning cannot do is ask the second question: what produced this error, and what does it reveal about how the system actually works? That question is the second loop. And it was the question that never got asked in Marcus’s shop — not because Marcus wasn’t smart enough to ask it, but because his training had never taught him that it was a question he was supposed to ask.

The salmon knows the river. It was built for that passage — every instinct calibrated to find the channel, fight the current, return to the place where the work of the species gets done. When something blocks the way, it tries again. Not because it is stubborn or foolish. Because it is doing exactly what it was built to do, as faithfully as it knows how.

It will not stop and ask why the passage is blocked. That question is outside the instruction set it was given.

Marcus was not the salmon. But the training infrastructure that produced Marcus — the degree programs, the CFRE curriculum, the AFP conference sessions, the practitioner literature — had handed him a very specific instruction set. Acknowledge. Steward. Repair. Try again. It was a good instruction set for a stable river.

What it did not include was the instrument for the second question.

Not: what went wrong?

But: what in our system produced this — and is it still producing it right now, in ways we haven’t noticed yet?

The donor Marcus lost wasn’t lost because he didn’t care. She was lost because the system that was supposed to carry her trust across a busy month had no mechanism for flagging when it was about to fail. The card went out late not once, but likely more than once, for donors whose relationships weren’t strong enough to survive it. Marcus never knew, because the single-loop correction — send the next one faster — made the symptom disappear without surfacing the pattern.

The Fundraising Effectiveness Project data has been telling this story for years. Four out of five first-time donors never give a second time. The sector raises more dollars from fewer people, year after year. The profession names this. It talks about it at conferences, writes about it in journals, builds sessions around it at chapter meetings.

And then it sends the next card faster.

Not out of indifference. Out of training. The profession was built — deliberately, carefully, by people who knew what they were doing — to fix the visible problem in front of it. It was never built to ask what produced the problem, and whether the same architecture is producing it somewhere else in the organization right now, quietly, in a place no one is watching.

There is a stretch of the Elwha River in Washington State where, for nearly a century, two dams blocked one of the most productive salmon runs on the Olympic Peninsula. The salmon tried. Every year, they found the base of the Elwha Dam and turned back. The river was right. The passage was wrong. And the salmon had no instrument for the difference.

In 2011, the dams began to come down — the largest dam removal project in American history. Within two years, salmon were swimming reaches of the Elwha they hadn’t accessed in a hundred years. The river hadn’t changed. The passage had.

The salmon didn’t remove the dam. They couldn’t ask why it was there. What they could do — what they did, faithfully, for a century — was show up. Keep trying. Make the case, in their way, that something upstream mattered enough to fight for.

The humans had to ask the second question.

In fundraising, we are the humans. We have the capacity the salmon doesn’t: to stop at the wall and ask what built it, and whether we have the tools to take it down.

Marcus does. He just hasn’t been handed them yet.

That’s what this platform is for.