There are two standard prescriptions for founder dependency. They are opposites, and they are both wrong.
The first one is *step back.* Delegate more. Get out of the weeds. This treats the founder’s involvement as the problem itself — so the founder pulls out of everything at once, judgment included. And then the thing that always happens, happens: quality drifts, a deal gets mishandled, something breaks. The founder watches it break. And their nervous system — the one from the last section, the one that keeps receipts — files the whole experiment as fresh evidence for the original rule. *See? Unwatched things break.* The grip comes back tighter than before, now with a recent citation.
Naive delegation doesn’t cure founder dependency. It manufactures the exact evidence that reinstates it.
The second prescription is the quiet one nobody says out loud: keep the founder in everything, because the founder’s judgment genuinely is better. And here’s the uncomfortable part — the premise is usually true. Twenty years of pattern recognition is real. The founder often *does* make the better call.
True premise. Wrong conclusion. Because it confuses two things that need to be pulled apart.
—
**Judgment** is the quality of the founder’s calls. The pattern recognition, the taste, the instinct for which client will be trouble and which risk is worth it. It’s genuinely valuable, and the goal was never to remove it from the company.
**Load-bearing** is something else entirely: it’s the founder’s *presence* being structurally required for routine things to move. Approvals queuing behind one calendar. Handoffs that only work through one inbox. Questions that can only be answered biologically.
Here’s the test, and it fits on an index card:
*Does this reach the founder because it’s genuinely hard — or because there’s no other way for it to move?*
If it reaches you because it’s hard, that’s judgment. That’s the job. If it reaches you because there is no other path, that’s load. That was never the job. That’s just where the weight ended up.
The problem was never that the founder’s judgment is in the company. The problem is that the judgment became the only connective tissue — the sole medium through which routine work coordinates itself.
—
A load-bearing wall and an architect both hold buildings up. They do it differently.
The wall holds the building up by standing there, forever, physically present, unable to leave without collapse. The architect holds the building up by having encoded their judgment into the structure — every beam placed according to decisions they made once, well, and no longer need to supervise.
Founder dependency is what happens when the architect gets bricked into the wall.
The way out is not less judgment. It’s *encoded* judgment. The decision criteria written down. The thresholds made explicit — what a good deal looks like, when to walk away, what gets escalated and what doesn’t and why. The instinct turned into teaching cases, the taste turned into standards. Slowly, deliberately: the founder’s judgment present in the system, without the founder’s body present in the loop.
That’s what actually retires the old rule from 4b, by the way. Not stepping back and hoping. Building the structure, watching it hold, letting the nervous system collect new receipts — this time in your favor.
Remove the load. Keep the judgment.
The company gets stronger on both counts. And the founder finally gets to do the job the title claimed all along.