The Context No Hire Can Carry

Before Jinn, every brand I ran kept its record in four places.

If that sounds like your brand, you're not behind. Four places is what a record looks like when nobody has been given the job of holding it in one, and at most brands nobody has.

The four didn't talk to each other. Nobody could read them in order, so nobody did. What a new person got instead was the tour: me, walking them through the brand, one session at a time, until enough of it had moved from my head into theirs that I could stop checking their work.

I've tried to put a number on what the tour costs. For any role, at any brand I've run, ten to twenty hours is the minimum. Realistically it's closer to a hundred: several sessions, sometimes weeks, sometimes months, before I had real confidence that someone understood the brand and, harder, how I wanted it operated. The quality of the documentation varied from brand to brand. The number didn't move much either way, because the documentation was never where the brand lived.

On a small team the tour has a second cost. Everybody wears several hats, the business depends on you staying productive, and the temptation is always sitting there: I can do this faster myself. Everyone knows that doesn't scale. Everyone does it anyway, because the alternative is another hundred hours.

The hire that can't be made

The obvious answer is to hire the operator. Someone senior enough to hold the whole brand, so the founder no longer has to give the tour. People assume I tried that and it went badly.

I never tried and failed. I don't think one person can carry the whole brand's context, and the reason has nothing to do with the people. I still work with agencies. I still hire. The people are the point, and the good ones are worth every hour. What every one of them was missing, the great people and the great agencies alike, was context, which is a different thing from talent. There is so much living inside a brand at any moment, a customer's complaint, the claim we pulled, why the packaging says what it says, which retailer wants what, that nobody outside the day-to-day can possibly know it, and nobody inside has time to write it down.

The second half is worse. Most people are deeply good at one thing, and that expertise doesn't translate. The person who understands your packaging doesn't know what the post-purchase email needs, and neither of them can tell you what the Amazon listing is doing to your search traffic. A brand is a set of decisions that influence each other, and the rarest hire in marketing is the one who can see how a decision in one function moves the others.

This isn't a small-company quirk, either. When Panopto and YouGov surveyed a thousand US employees at companies with at least two hundred people, 42 percent of institutional knowledge turned out to be unique to the individual, held by one person and no one else, and the average employee spent 5.3 hours a week waiting on a coworker for assistance or insight (Panopto Workplace Knowledge and Productivity Report, with YouGov, 2018). The survey stops at companies of two hundred people, so take the six-person brand as my extrapolation, not their finding: the same knowledge in one head, with fewer people to wait on.

When the knowledge walked out

The version of this I lived closest was in one of my other businesses. A partner left, for personal reasons. We're on good terms and still close friends, and what follows isn't about the leaving. It's about what was being carried. My partner had run big portions of the marketing and a lot of the manual work, and had done it well, in a way that was entirely personal: a process and a technique built over a long working relationship, the kind no job description captures and no handover meeting transfers. I picked it up in passing, because that's the only way anyone picks up work like that, and replacing it with another operator turned out to be very hard. Good operators aren't scarce. What needed replacing was a way of working that had lived in one head.

The timing mattered. This was the moment the language models got genuinely good, and the amount of context they could hold at once started expanding fast. So I experimented. Automate one piece of what had walked out the door. Then another. Then wire the pieces together until the brands ran without that involvement. It worked well enough to convince me of something I hadn't believed a year earlier: that AI could be part of how a small brand holds itself together. And it settled something else. I didn't want the business carrying that risk again, because if somebody leaves, a big portion of the knowledge goes out the door with them.

The brain was portable

Jinn didn't start as a brand record. It started with video, the hardest thing I wanted to make. Video punishes every gap in the brief. Wrong claim, wrong color, wrong tone, and it's wrong in motion, in public.

I was building it for myself, so I wasn't willing to sacrifice quality to get something out, and the brain behind the video got very detailed and very complicated. I probably overbuilt it. Then I noticed something about the overbuild: the brain was portable. The same record that could brief a video could brief an ad. It could brief an email, a listing, a strategy document. Every output I'd ever paid someone to make needed the same context, and here it was, in one place, finally organized well enough that something other than me could read it.

That was the click. What I'd been trying to hire was a brain that every hand could read from, and I'd built one by accident, for video.

Alive, or it's a PDF

What a record like that has to be, above everything, is alive. A PDF is a photograph. It was true the day it was taken. Brands move, categories move, the platforms move under you, and the record has to move with them, which was very hard to do by hand and is exactly why every brand book I ever commissioned went stale within a year. So the record gets fed constantly. It remembers. It learns from feedback the way a good hire does, and as the brand changes it changes too.

Why the brand never got a system of record the way money and customers did is its own piece. The operator's side of that argument is simple: I can constantly feed the thing what I know, as long as it's organized properly and can learn from me, and it never needs the tour again.

The mistake I learned from

There's a mistake we made at Positive Pretzels that I still think about, and it's connected to all of this.

We were careful. We were unwilling to scale as quickly as we could have, because of what we saw as reputational risk: put something out that isn't quite right and the brand pays for it. We were wrong, and being wrong that way cost more than the risk we were protecting against. Everyone is busy, the volume of everything is enormous, and consumers are more forgiving than a careful founder assumes. Moving fast is how you find what resonates, and course-correcting sooner is how you learn cheaper. I believed that then. What happened to that business is why I hold it as a rule now.

What made the caution feel rational at the time is that every piece we put out had to pass through the one head that held the context, because nothing else could check it against the brand. Speed and safety were in tension only because the context couldn't travel. Once it can, once the record itself can check the claim and the voice and the color before a person ever looks, the tension goes with it.

Who builds a thing like this

The team behind Jinn is lifelong marketers, career technologists, and people who have been investors, strategists, and CPG executives, building one thing together. Two of them are co-founders I've known for more than ten years: Aaron Hassell, who has been helping develop the underlying product, and Sarah Aylward, who is leading marketing and growth. That mix is the cross-function problem from earlier, the packaging person who can't see the retention email, solved at the team level before it was solved in the product. A single-discipline team could not have built it, for the same reason a single hire could not have carried it.

What I stopped doing is asking the people I work with to carry the context, because that was the one thing the tour never got across and the one thing the business couldn't afford to lose again. The operator I always wished I could hire turned out to be a memory, one that every person I work with gets to read from on their first day.