I Launched It Without Permission

Two months ago I launched a PLG campaign without asking my CEO.

Fifty-one accounts signed up. Thirty-seven are onboarded and consuming right now. That happened in eight weeks, from a motion the company had been told wasn't the priority.

Three weeks after I launched it, I took the CEO seat.

I want to be precise about that order, because "founder goes rogue" is a story people enjoy for the wrong reasons. I didn't launch it because I had the title. I got the title because I launched it.

This isn't a story about a bad CEO. He was running the plan we had agreed on, and running it competently. And I want to kill one version of this story before anyone tells it for me: I didn't sit there nodding. I fought. I made the argument for this motion more times than I can count, in rooms where I was the most informed person present.

I just made it without a vote. I had given that away.

Hand-to-hand combat doesn't compound

The motion I went around was hand-to-hand combat. Every deal a bespoke conversation, every conversation a custom explanation, every customer landed by hand.

That motion works. It closed real business and it taught us things about buyers that no signup form would have. What it doesn't do is compound. You can add people to it, but you can't get leverage out of it, and the day you stop pushing is the day it stops.

Here's the part I own, and it isn't the plan. Selling one at a time means nobody ever has to commit, in public, to a single sentence about what the product is. I said that at the time. I said it repeatedly. What I had done, well before any of those conversations, was hand over the authority to act on it, and an argument you can't close is just a position you keep restating.

So the motion stayed. Not because anyone overruled a good idea out of stubbornness, but because the org chart I built made my conviction advisory. That's on me, and it's a structural mistake, not a character one. Those are the expensive kind, because they keep costing you while everybody involved is behaving reasonably.

How I got there

I started this company on my own. Then I decided I wasn't qualified to run it.

I could build the product. I'd spent my career on the marketing side of the data problem, close enough to engineering to know what was possible and close enough to the buying to know what was useless. What I'd never done was run a company. So I brought in leadership to do that, and the title went with the job. I handed it over without much ceremony, because I read the title as a description of a skill set I didn't have.

Some of what I learned in those two years I couldn't have learned any other way, and the people who taught me were good at their jobs. They brought operating discipline to a company that had none. They built process I still run on. They executed the plan we agreed on. I'm not writing the version of this story where they're the villain, because that version isn't true and it would be a cheap way to make myself look better.

What is true is that I let the company get devalued, and that's mine. Not because I went quiet. Because I structured things so that being right didn't settle anything. I kept losing arguments I should have been ending, and I lost them on authority rather than on merit, which is a thing you can only do to yourself.

That's the actual lesson and it took me two years to learn it. Conviction without decision rights is just commentary. You can be the most informed person in the building and it buys you nothing if you've arranged the company so that your conclusion is one input among several.

What I was arguing about

Eleven inventions. The provisional is filed and we're working through the individual applications now.

I'm careful with that wording. "We have eleven patents" is exactly the kind of sentence this industry says without blinking, and it's the kind I'd catch someone else on. We don't have eleven grants. We have eleven things worth patenting, filed, with the real applications in flight. In a market this loose, that's the rarer claim anyway.

We are not a rebranded LLM with a nice landing page. We're the only company that can set identity or deploy a customer tag carrying identity. That's the whole thing. That's the primitive everything else in this market is downstream of, and most of what's being sold around us is slop built on top of somebody else's graph.

I knew that the entire time. I said it constantly. It still got positioned as a feature, because the person saying it wasn't the person deciding, and I was the one who had set it up that way.

That's the part I'd undo if I could. Not the hires. The handover.

The pricing model was the same disease

Watch what happened to our pricing and you can see it from the outside.

We started usage-based. That was right. We run a graph. What a customer gets is a function of how much of it they draw on. Every resolution, every lookup. There's no version of this business where the value delivered and the money charged move independently.

Then we became a SaaS company. Not because a customer asked. Because that's what a company like ours is supposed to look like on a slide. Seats, tiers, platform fees, annual commitments. The shape investors recognize, the shape that gets a multiple. Every argument for it was a good argument, and not one of them was about the product.

We're back to straight usage-based now, where we started. Customers pay for what they take. The graph gets better, they take more, we make more. We stop earning it, they stop paying.

Round trip cost us time. What it bought is that I hold this position now because I ran the alternative, not because it feels right.

I built this for myself

Here's the recalibration.

I didn't build Delivr for a persona I researched. I built it for who I already was: a technical, data-driven marketer who understands the science and the data. Someone who's been handed a resolution rate with no way to verify it. Who's been sold "intent" that's a black box with a confidence score painted on the side. Who's had to answer "who was on our site" with a number he couldn't defend to his own CFO.

That guy is the customer. And when you've been the customer, you have an instrument nobody else in the building has. You can tell in seconds whether a feature is real or whether it's a slide.

For years I discounted that, because "I would want this" sounds like the least rigorous argument in the room. It's the most rigorous one available, as long as it's true. The trap is founders who say it and were never the user. If you were, that's not a preference. It's compressed expertise, and it moves faster than any process you could build to replace it.

What trusting your gut actually means

Trusting your gut isn't a license to ignore data. I run a data company. It would be an ignorant thing to argue.

It's narrower than that. Data tells you what happened. Your gut tells you which question to ask next, and which of the forty defensible options is the one you'll still believe in a year. There's no dashboard for that.

My test: if I disagree with the room and can't say why, I wait. If I disagree with the room and can name the exact customer moment they've never had, I go.

The PLG launch was the second kind. I'd sat in the seat of the person who was going to sign up. I knew what he'd do when he hit the page. Fifty-one accounts later, that turned out to be worth more than the meeting where I'd have asked permission.

Execution

Conviction without follow-through is just an opinion, so here's the other half.

One filter runs the roadmap: does this make the graph better, or does it make the deck better? Identity comes from the graph. Intent comes from identity. Everything we sell is downstream of whether those resolutions are correct, and nothing downstream compensates when they aren't. Deck items aren't evil. They're just not the work.

I hire differently now, too. Not better people. A different brief. I used to hire people to tell me what the company should be. Now I hire people to make what we've decided real, and to fight me hard before the decision instead of relitigating it after. That's a different job and it asks for a different person. The good ones don't want a founder who outsources conviction. They want to know what we believe so they can go be excellent inside it.

The team is the reason this worked

I've written this as though the last two months were a decision I made. The decision was mine. The result isn't.

Fifty-one signups and thirty-seven live accounts is not something a founder does. It's what happens when the people around him stop waiting to be assigned work. That's what changed here, and I didn't engineer it. Once the direction got clear, this team took real ownership of it. Not "aligned with the strategy." Owned it. People made calls in their own lane that I never saw until they were already working, and they were right, and the thing shipped faster than it would have if it had come through me.

They are one hundred percent the reason we're seeing what we're seeing. I get to write the essay. They did the work.

The other reason I drive this hard is the investors who put their own money into this. That's not an abstraction to me. Somebody earned that, decided this was worth the risk, and handed it over on the strength of what I told them we'd build. I think about that more than I think about the cap table. You do not get to be casual with money somebody worked for.

Between a team that owns the outcome and investors who trusted me before there was proof, the only unacceptable thing left was me holding back.

The actual path

I've held the CEO title for six weeks. It changed less than you'd expect, because I'd started doing the work three weeks before anyone made it official. That's the part worth keeping. The title didn't authorize the decision. The decision earned the title.

Nobody hands you this job. You start doing it, and the paperwork catches up.

If you're a founder who's talked yourself into believing the operational parts of the job disqualify you from leading it: bring in the help, learn the craft, take it seriously. Then take the wheel back. The company you built for yourself only works when the person it was built for is driving.

We're hiring and we're building the graph out fast. If you want to know what we believe before you talk to us, you just read it.