In August he took two weeks off for the first time in nine years. Eighty staff, services-led, profitable. By day four his phone had three escalations he had to make himself. Two were standard discount calls he’d handled a hundred times. He came back exhausted, mostly at himself.
His senior team are good. They couldn’t step in because they didn’t know what they were allowed to step in on. The answer to “what would Dave decide here” only lived in Dave. That’s the trap. The infrastructure for someone else to act in his place had never been built.
What does founder dependency actually look like?
Founder dependency in a services business shows up as the team escalating decisions you wish they’d handle, every time. Senior people ask permission rather than acting. Clients ask for you specifically. The phone doesn’t stop on holiday. The pattern is consistent across every business I’ve worked with where the founder is still in every decision. The visible symptoms come from the same structural cause.
The structural cause has three layers, usually all present.
Decisions route back to you because the team has no clear guidance on what they can decide alone. Without a written line, the safe move is to ask. The team isn’t being timid; they’re being responsible.
Judgement isn’t captured anywhere. Even when the team can decide, they don’t know how you’d weigh the trade-offs. So they ask, because the wrong call lands harder than the slower call.
Decision rights aren’t defined. The boundary between team authority and founder authority is implicit, which means in practice it gets re-litigated on every edge case. The founder ends up adjudicating the same boundary disputes repeatedly.
All three compound. Fixing only one doesn’t break the trap. The diagnostic Legal and General published a few years ago captures the consequence. More than half of UK small businesses say they would stop trading within twelve months if they lost an owner, director, partner, or key person. The founder is the system, and the system has one node.
Why does delegation advice keep failing?
Most delegation advice tells founders to trust the team, step back, and get out of the way. It assumes the team can step in if the founder makes the space. In a structurally founder-dependent business that assumption fails. There’s no decision-rights map, no captured judgement, no defined boundary. Stepping back without first building any of those just leaves a vacuum, and the vacuum routes straight back to the founder.
The advice fails for a specific reason. The team isn’t underqualified. The judgement they need to apply hasn’t been written down. So when they’re left to decide, they reach for the safest option, which is to escalate. Or they make the call and live with the anxiety that the founder would have done it differently. Neither produces what the founder wants, and neither is fixable by exhortation.
Founders typically respond by trying harder. More one-on-ones. More talking through the trade-offs. More “I trust your judgement, just call it”. Each helps in the moment. None of them stick, because the same conversation has to happen again on the next edge case, with no permanent artefact left behind. Six months later the founder is having the same talk with the same senior person about the same kind of decision. That’s a sign that the conversation is the wrong artefact for the work.
The right artefact is a structure the team can read without you in the room. Once it exists, the conversation changes from “what would Dave do” to “here’s how we decide this”. The work compounds for the team only when the judgement is in something durable enough to survive a holiday.
Why services businesses get the worst version of this
Services businesses get founder dependency worst because the founder often is the system. In a product business, processes substitute for judgement. In a services business, the call on whether to give a discount, scope a project, fire a client, or hire a senior person reflects the founder’s individual taste. That taste is the product. Until it’s externalised, the team has nothing to substitute for it.
In product land, the founder writes a process for assembly, ships it to the floor, and the floor produces the output without the founder there for each unit. In services land, every client interaction is a small judgement call. How firm to be on scope creep. How much to flex on a deadline. Whether the project gone sideways needs the founder in the room. The judgement is constant, and it’s harder to write down than a process.
Then there’s what the agency-owner literature calls the named-door problem. Clients buy you specifically. They want the person with their name on the door. The team can do the work, but the relationship sits with the founder, and the relationship is half the value.
The combination is what makes the services version of the trap hardest to escape. The infrastructure work is harder, the substitutes are weaker, and the cost of skipping it is higher. The honest version of the work is to externalise the judgement first, separately from the relationship work. Client relationship transition is a different project, longer-running, that runs after the daily friction is broken.
What infrastructure actually breaks the trap?
The infrastructure that breaks founder dependency has three pieces. A decision rights map: a one-page document showing what the team can do alone, what needs a check-in with you, and what needs your direct call, with examples in each column. A written description of how you weigh the most common high-stakes calls. A simple structure for surfacing the exceptions the map didn’t cover.
None of these are exotic. The decision rights map is a Google Doc with three columns and concrete examples. The written judgement is your own thinking on the half-dozen calls you make most often, captured in plain language. The exception surfacing is a fortnightly thirty-minute slot where the team flags the calls they had to make that didn’t fit the map.
A worked example helps. The team can refund up to ten percent of an invoice without asking. They can give a discount up to fifteen percent for a multi-month commitment without asking. Discounts above twenty-five percent come to you. Scope changes within a project can be agreed by the project lead if they don’t extend the timeline by more than two weeks. Anything beyond that comes to you. The numbers reflect the founder’s actual tolerance, written in language the team can read at speed.
The honest cost of building this is six to twelve months of disciplined work, often during a stretch when you don’t feel you have the time. The urgent crowds out the important, and founder dependency makes the urgent more urgent. Most founders skip the work because they can’t carve out the time. Visibility into the value isn’t the issue.
Once it’s built, every additional decision the team makes without you is a unit of capacity returned, a unit of judgement compounding inside the team, and a unit of risk taken off the L&G stat. The next August is different. The phone is quiet. Two of the calls that would have come to you got handled at the right level by the right person, the way you would have made them, without you in the room.
That’s what freedom in this kind of business actually looks like, in unglamorous detail. Book a conversation if any of this sounds like the August you’ve just had, or the August you’re about to have.
