Positive Pretzels made the first organic flavored hard pretzel in the United States, and we were fortunate enough to get picked up by Whole Foods in the Midwest. I want to be exact about what that did for us, because it wasn't just the volume.
It was the sentence. "We're in Whole Foods." Say that to a distributor and a door opens that was shut the month before. Say it to a customer and they trust the product before they've tasted it, because a retailer known for reading ingredient panels has said yes to ours. The credibility did work that none of our own messaging could have done, and it did that work everywhere else we sold. It helped the rest of our distribution fall into place. It got us into conversations we would never have been invited into otherwise. A small brand spends most of its first years trying to borrow trust from somewhere, and a shelf in the right store is the cheapest trust I ever borrowed.
Then there was the shelf itself. Where a product sits decides whether it turns. Hand level and eye level get picked up and sell through better than the bottom shelf, and I ran enough of my own studies on that to stop arguing about it. So we would go into stores and fight for the planograms, the diagram that says which product goes on which shelf, at which height, with how many facings. A planogram is a negotiation, and you fight it in person, one store at a time, with a store manager who has forty other brands asking for the same six inches.
A small brand spends most of its first years trying to borrow trust from somewhere, and a shelf in the right store is the cheapest trust I ever borrowed.
That was the first shelf I fought for. Watching the shelf move three times since, from wood and metal to a results page to a product page, is a story I've told on its own, and I won't retell it here. This piece is about what I learned on each of the ones I stood in front of, and about the fourth one, which nobody can stand in front of at all.
The words that mattered
Dandy, the pet-supplement brand I ran, was truly personalized, thousands of combinations, and the customer who wanted that wasn't walking a store aisle. They were typing.
Search is high-intent and close to the bottom of the funnel. A person searching for a personalized supplement for their dog is most of the way to buying one, and the only question is whose page they land on. So the energy at Dandy went into ranking for the words that mattered, so that the people looking for exactly what we made could find us. It was a very important channel for us, and it taught me something the shelf never had to. On a shelf, the buyer discovers you. In search, the buyer arrives already knowing what they want and you have to be standing on the exact words they use to ask for it, which is a narrower target and a much better one, because the person who finds you that way is already sold on the category and only deciding on the brand.
Even so, it had the same shape as the planogram fight: get in front of the buyer at the moment they're choosing, in the place where they choose. The shelf had become a results page, and eye level had become page one, and I recognized the fight even though the store had disappeared.
The tipping point
I've built brands on Amazon too, and Amazon is where the fundamentals stopped being a feeling.
Amazon is a data company, and the king of optimizing for profit. To operate well there you look at share of purchase, rankings, conversion, and you know Amazon is looking at the same numbers you are. You learn how the game is played and what Amazon wants from a listing. You can see where you're winning on a keyword and where you're losing, and where a competitor is winning instead of you, and by how much. It got super analytical, and for me it was a tipping point. For years I had known things about placement the way an operator knows them, from standing in stores and watching what moved. Amazon was the first shelf where I could say the second half of the sentence: sure, we knew this was right. Now it's statistically provable as right.
| Known from standing in stores | Statistically provable as right | |
|---|---|---|
| How I knew what placement did | Years of watching what moved on a shelf, plus enough of my own studies on hand and eye level to stop arguing about it. | Share of purchase, rankings and conversion, keyword by keyword, with the competitor beating me on that word visible beside my own number. |
That sentence is a large part of why Jinn exists. Brand-building sits in a space between two things. On one side, creativity and messaging, the things that are genuinely hard to define and that make a brand a brand. On the other, true, quantifiable analysis. Retail lived almost entirely on the first side, with velocity reports to keep it honest. Amazon dragged a big piece of it to the second side and left it there. I've spent every year since trying to hold both at once.
The same fundamentals
When the fourth move started, the first thing I told my own team was that it's really the same fundamentals.
A buyer asks something like "I need a supplement for my dog," and the thing they ask is ChatGPT, used as a super intelligent Google. Back comes a paragraph with names in it. That paragraph is a shelf. It has a top position and a bottom position and a large number of products that aren't on it at all, and the questions I ask of it are the ones I asked of every shelf before it. Are we being recommended? What are they saying about us, and is it even right? We built Fama to ask those two questions on a schedule; the mechanics have their own introduction, so I'll leave them there.
Wood and metal, a planogram to argue over, and a store manager with forty other brands asking for the same six inches. A door I could walk through, one store at a time.
The buyer arrives already knowing what they want, so you stand on the exact words they use to ask for it. Still a door: page one is a place you can work toward, word by word.
Change a listing, watch share of purchase and conversion move, and try again. Amazon never let me argue with it either, but that is a door of a different kind, and I got used to it.
No store to drive to. The planogram does not exist, and neither does the keyword report where I could see exactly which word we lost, and to whom. A paragraph, not a place.
I think being accurately described, described on-brand, and visible inside the answer is going to be the future for all of e-commerce, and honestly for anybody doing business online. I hold that view the same way I held the search view at Dandy, early and on the strength of watching where the buyers went.
And check what the shelf is saying about you, because for the first time in my career, the shelf talks.
What I want to dwell on here is the one thing the fourth shelf changed, because it's the thing I'm still getting used to. I can't walk into it. There's no store to drive to. The planogram I used to argue over doesn't exist, and neither does the keyword report where I could see exactly which word we lost, and to whom. When the answer is wrong about your brand, you can't correct it in place, and you can't necessarily see where the wrong fact came from. On the first three shelves the fight was hard but the door was there. On this one the fight is just as real and the door I used to walk through doesn't exist.
I'll admit the part of me that fought for six inches on a planogram finds that unsettling. The Amazon years make it feel familiar, because Amazon never let me argue with it either. It let me change a listing, watch share of purchase and conversion move, and try again, which is a door of a different kind, and I got used to it.
The catch is that this shelf will describe you whether or not you show up to it.
What the package used to carry
The part of the retail years I think about most is the package.
For a retail brand, so much is in the packaging and so much is in the messaging on it. You have a few square inches. In them, a buyer has to get what the product is, sense that it was made with care, feel that it's for them, and decide that it's trustworthy. All of that had to fit on the box, because for most buyers the box was the only narrative they would ever get from us. It had to do that work alone, with nobody from the brand standing next to it, in a store we might visit twice a year, in the moment a shopper gives a package before their hand moves on.
| The package | The AI answer | |
|---|---|---|
| What it can carry | A few square inches, and one narrative for everyone who picks it up, with nobody from the brand standing next to it. | One true thing for the person asking about a supplement for their dog, and a different true thing for the person asking whether to switch toothpaste. |
The fourth shelf is the first one where that's no longer true. When the engines describe your brand, the narrative isn't tied to a package anymore. It can be shaped. It can carry more of the story than a box ever could, and it can say one true thing to the person asking about their dog and a different true thing to the person asking whether to switch toothpaste. That's more reach, and more personalization of the message, than a package could ever hold. It's the upside, and it's the reason I want to fight for this shelf the way I fought for hand level.
The catch is that this shelf will describe you whether or not you show up to it. On a wood shelf, absence was just absence: the product wasn't there, and nobody said anything about it. On the fourth shelf, absence gets filled in by whatever the engine can find, and some of what it finds is wrong, and it's said in the same confident voice as everything else. So the fundamentals hold, with one addition. Get in front of the buyer where they choose. Fight for the position. And check what the shelf is saying about you, because for the first time in my career, the shelf talks.
What is the fourth shelf?
The paragraph an AI engine returns when a buyer asks it what to buy. It has a top position, a bottom position, and a large number of products that are not on it at all, which is what makes it a shelf.
How is it different from the three shelves before it?
You cannot walk into it. There is no store to drive to, and the planogram you used to argue over does not exist. When the answer is wrong about your brand, you cannot correct it in place, and you cannot necessarily see where the wrong fact came from.
What happens to a brand that ignores it?
The shelf describes the brand anyway. On a wood shelf, absence was just absence: the product was not there, and nobody said anything about it. On this one, absence gets filled in by whatever the engine can find, some of what it finds is wrong, and it is said in the same confident voice as everything else.