The Phantom Org

Your org chart shows one company. Your actual operating reality is another. You know this. Most of the people around you know it too. Nobody says it out loud.

There is a function — probably more than one — where the title belongs to someone else, but the decisions still come to you. You review the thing. You edit the thing. You catch the thing before it goes out. You tell yourself you’re “just staying close” or “helping them ramp.” What’s actually happening is that you have a VP in name and a VP-flavored assistant in practice.

The org that exists on paper

Walk the org chart from your last board deck. Five functions, five owners. Now trace where the final calls actually land. The marketing copy that went out last week — who approved it? The pricing exception on the mid-market deal — who said yes? The offer for the senior engineer — who set the number? If the honest answer keeps coming back me, you’re running a phantom org: titles on paper, decisions in your inbox.

Your team has already adjusted to it, which is why it’s quiet. They pre-clear things. They wait. They’ve learned that final calls come back to you eventually, so they’ve stopped making them — not out of weakness, but out of accurate pattern recognition. The org chart says delegation. The Slack threads say otherwise.

You’ve felt the private version. It’s 11 p.m. and you’re rewriting a two-line Slack reply to your own VP for the third time — softening it, because what you want to type is this isn’t the call I’d have made, and you know what sending that does. So you soften it. Then you take the decision back anyway. Quietly.

Why founders rebuild the cage

Noam Wasserman’s Founder’s Dilemmas names why this happens: founders consistently under-hire and over-review because the perceived cost of a mistake feels bigger than the actual cost of staying involved. The mistake is vivid — a botched launch, a bad hire, a lost deal. The cost of staying involved is invisible, because it’s distributed across every decision your team didn’t make, every week, indefinitely.

Ben Horowitz goes further — he calls the rehire phenomenon specifically: you bring someone in, quietly take parts of the work back, and end up with the worst of both worlds. You’ve paid for a senior hire and you’re still doing the work. Run the six-month test on your most recent senior hire. Did they take it, and you haven’t looked back? Did they take it, but you’re still checking in more than you expected? Or did they take it — and you’ve quietly taken parts of it back?

And underneath all of it, the question nobody asks out loud: if someone else can really run sales — run it well, without you — then what, exactly, do you do here? That question doesn’t get asked in any meeting. It runs the org chart anyway.

First Round’s research on the $1M→$10M transition is blunt about the stakes: the founders still holding shadow functions at that band are the ones who plateau at that band for years.

How to know it’s you

Count the Slack threads pinging you right now, waiting on your input to move forward. A few real ones is a weekday. More than you can clear today — and that’s been true for months — is a structure, not a workload. Then there’s the hire you’ve been circling and haven’t made. What’s the real reason? Sometimes it’s genuinely timing. Sometimes it’s cost against an unknown return. But if you sit with it and the honest answer is closer to what it would mean to hand that over — that’s the pattern talking.

And the question you’d rather not answer: walk the chart and count the functions where, if pressed, you’d admit you’re still running the show. “I’d rather not count” is a count.

What it compounds into

Shadow functions rarely exist in isolation. Often you hold them because nobody else knows the story well enough to make the calls without you — new hires ask “what should we do about X?” and the honest answer is “it depends on who we’re for,” which is the question that hasn’t been answered clearly. Without that clarity, every delegation is conditional, and every senior hire has to read your mind to succeed. Your team isn’t underperforming. They’re operating without the document they’d need to perform.

The functions you hold longest are the ones closest to revenue — sales, marketing, maybe customer success — because if they broke, the company wouldn’t survive the quarter. That’s real. It’s also exactly the argument that prevents you from ever letting go, because those functions will always be closest to revenue. Waiting until it feels safe to delegate revenue-adjacent work is waiting for a moment that never comes.

The hand-off happens while it’s still scary.

Meanwhile your week fills with functional work — approving, reviewing, unblocking — and the CEO work goes undone. The calendar doesn’t show it, because Slack threads don’t appear on calendars. But it’s where most of your week goes, and it’s its own pattern.

Naming the functions: three moves

Take the honest census. List every function. Mark the ones where the title and the decisions diverge. Then mark the one or two you should still hold at this stage — naming those out loud is what makes releasing the rest credible.

Write the mandate you never gave. Pick one shadow function. One page: what the owner decides without you, what they bring to you, what good looks like in ninety days. Most failed delegations fail here — the hire got a title and a backlog, never a mandate. Hand the function over with the document.

Move yourself from pre-approval to post-review. Stop being the gate that decisions wait at. Review outcomes on a cadence instead of approving inputs in the moment. Your team will make calls you wouldn’t have made. That’s not the risk. That’s the point.

Name the functions. Some you should still hold. Most you shouldn’t. The first move is honesty about which is which.