On November 20, 2025, a product called Team-GPT had one million impressions and two thousand clicks on Google per month. The next day, it became Juma and six weeks later, the numbers fell to 90% in reach. An 83% collapse in clicks.
Iliya Valchanov, the founder, posted the story publicly. Thanks to founders like him who share these cases openly, we get to deconstruct and analyze not only the technical SEO side — and I highly recommend reading the comments on that post — but also the positioning and strategic layer underneath.
Iliya isn't failing here. If anything, he's nailing one of the hardest challenges a founder can face — a full-stack rebrand at scale, executed with discipline and public transparency. But nailing the migration doesn't cancel the bill. The tax isn't for how you leave borrowed ground. It's for having built there in the first place. And that's what this essay is about: the cost of building on borrowed ground.
How Founders Use Borrowed Positioning to Get Early Traction
Rewind to 2023. ChatGPT is the fastest-growing product in history. "ChatGPT for teams" becomes a search query before anyone has built the thing to answer it. And a founder realizes something: put those three letters in your product name, and Google does half your marketing for you.
He calls the product Team-GPT. The wedge works immediately. People typing "ChatGPT for teams" land on his page. The brand rides distribution it hasn't earned by itself, and over the next two years that distribution produces a million impressions and two thousand clicks a month — real traction from a clever positioning bet.
This is the part founders get wrong when they moralize about positioning. Borrowed ground is not a character flaw. It is a legitimate strategy in the zero-to-one phase. You plug a gap next to a dominant platform. You rank for adjacent intent. You live in someone else's gravitational field long enough to build a product and a cash flow. Many great companies start exactly this way.
Airbnb in its early days, trode Craigslist's distribution. They built tooling that let hosts cross-post their Airbnb listings to Craigslist with one click, siphoning renters off the biggest housing marketplace on the internet. It was clever, borderline gray-hat, and it worked. They grew on Craigslist's gravity until they had enough two-sided liquidity to stand on their own name. Then they walked away from the borrowed ground before Craigslist could close the door.
The mistake isn't starting there. The mistake is staying.
Why Rebranding Destroys Your Organic Traffic
The deeper you build on borrowed ground, the more expensive it becomes to leave. And you don't get to decide when you leave.
Sooner or later, the platform you’re piggybacking on closes the gap and ships the native feature your wrapper depends on. Or Google reshuffles its algorithm. Or—like Team‑GPT discovered—the lawyers arrive. One day, OpenAI’s legal team reached out: the GPT trademark was theirs, and a product called Team‑GPT was standing in a field they had every right to clear. Conversations like that end only two ways—you rebrand, or you spend your days in court paying for the privilege to stay.
Iliya chose to rebrand. That was the only real choice available, but it was still a choice about how to leave — and the how is where founders either handle it with grace or lose the plot. A weaker response would have been to drag out the transition, half-rebrand, keep redirect hacks running, try to preserve the old SEO footprint through tricks. He didn't. He made it clean: new name, new domain, full migration, public transparency about the cost. Juma — a name that could mean what he wanted it to mean rather than a fragment of someone else's.
And the ground took 95% of his organic distribution with it, because it was never really his to begin with.
That's the part of the story that matters. The collapse wasn't a failure of execution. He did the technical work: 301 redirects, Change of Address in Search Console, on-page SEO, off-page link building, PR and podcast appearances to rebuild the entity. The collapse happened despite all of that, because Google had been sending traffic to the borrowed name, not to the company underneath it. Remove the borrowed name and the traffic has nothing to attach to until the new name earns its own recognition. Recovery usually takes six to nine months.
That is the exit tax. And it is survivable — Juma is rebuilding — but it is also non-negotiable. You either pay it when you leave, or you pay a different tax by staying.
Why the Real Cost of Rebranding Is Narrative, Not Traffic
Here is the part most founders miss. The cost of borrowed ground isn't primarily legal or technical. It isn't even primarily traffic. It is narrative.
When you build on someone else's name, the market files you under a story. When you move — or when the platform moves under you — that story either lets go cleanly or it marks you.
Iliya's story lets go cleanly, or will, given time. He walked toward something rather than away from something. He paid the tax visibly. He is doing the slow, unglamorous work of rebuilding entity recognition the hard way. The narrative that attaches to Juma won't be they used to be Team-GPT. It will be they grew up.
Contrast that with what happens when the borrowed narrative becomes the identity. There was a product called ClawdBot — built on Claude's API, lifting not just the infrastructure but the phonetic shape of the brand, the claw imagery, the full gravitational field of a platform it didn't own. It rebranded twice, each time pulling the borrowed narrative a little tighter rather than building something that stood on its own ground.
When Anthropic announced a remote desktop feature, the founder responded as though a tenant had been told the landlord was renovating — publicly, with visible frustration, as though the update were being taken from him rather than added by the people who built the platform. The API access was shut off shortly after. The market watched all of it. The narrative that formed wasn't about the product. It was about a persona that had confused someone else's momentum for its own.
He eventually went to work at OpenAI, renaming the project OpenClaw on the way out. I'm not judging that exit. Borrowed ground forces a reckoning and everyone crosses it differently. But the narrative tax doesn't dissolve when you leave. It follows the person.
From Borrowed Ground to Owned Demand
The Machado line — caminante, no hay camino, se hace camino al andar — comes from a Spanish poet, but the Menahan Street Band turned it into something else: a feeling. Your road doesn't exist yet. It appears as you walk it.
Borrowed ground feels like a road. It has pavement. Other people have walked it. You can move fast on it. But it was built for someone else's destination, and it ends wherever they decide it ends.
The founders I've watched escape this trap didn't do it by being more ethical than everyone else. They did it by getting honest about what they were trying to own. Once you know what demand you're actually building toward — not borrowing, owning — the positioning question gets simpler. What can we say that is ours? What can we stand on that doesn't require someone else's gravity to stay upright?
Owned demand is slow to build and almost impossible to steal. You can't algorithm-update it away. You can't ship a native feature that replaces it. It accumulates with every kept promise, every clear position, every time you hold the line when it would have been easier to borrow someone else's momentum.
The 95% traffic drop is survivable if you have something underneath it. A real audience that came because of you. A position that's yours. A narrative that's clean. That is what Iliya is betting on, and based on how he's handling the transition, it is a bet worth making.
If your positioning depends on someone else's gravity to hold its shape, that is worth knowing now — before the floor moves, or before you decide to move off it yourself.
