We operate a small industrial goods seller — European ownership, a US company, a commodity product sold both on Amazon and direct. Last month we pulled its channel data apart and found something that has changed how we argue about branding.
One channel converts at 17 to 18 percent. Every other channel it runs converts at 0.21 percent, and needs roughly 500 sessions to produce a single order.
Same product. Same price. Same warehouse. The difference is whether the visitor arrived already knowing what they wanted and who they were buying it from.
That is what branding is worth, stated as a number: the difference between needing 500 visits to make a sale and needing six. Not awareness, not sentiment — order economics.
Most people who register a US LLC and start selling on Amazon are optimizing the 0.21% channel forever. They get good at it. Some get rich at it. And they remain structurally unable to build the other one, because everything that produces a 17% conversion rate is an asset they do not own.
What you actually own on a marketplace
You own inventory. That is the complete list.
You do not own the customer relationship — the buyer is Amazon's, and you are not permitted to contact them outside the platform. You do not own the storefront URL, the reviews, the ranking, or the right to keep selling next quarter. You do not own the traffic: it was rented per unit, and it stops the moment you stop paying or the moment the algorithm reweights.
None of that is a criticism of Amazon. It is a description of the deal, and for many sellers it is a good deal. But it explains why so many profitable sellers feel fragile: a business with one channel it does not control is not diversified by revenue, it is exposed by structure.
The seller we run made one small move that illustrates the whole transition. In June it removed every mention of Amazon from its own website. Fulfilment still runs through Amazon's logistics — that part is operational and invisible to the buyer — but the site stopped presenting itself as a marketplace listing and started presenting itself as the manufacturer's home. Nothing about the product changed. What changed was which entity the customer believed they were dealing with.
The first asset: an entity that actually owns the brand
Before a website, before a logo, before any of the work people think of as branding, there is a legal question almost every seller gets wrong.
Who owns the trademark?
In the business we operate, the mark is registered to an individual — not to the company that sells the product. This is extraordinarily common and almost always accidental: the founder files the application in their own name because they are the one filling in the form, and nobody revisits it.
It matters more than it sounds. A brand owned personally rather than corporately is not on the company's balance sheet, complicates any future sale or investment, creates a licensing question between founder and company that nobody has documented, and can separate the mark from the business entirely if ownership changes. You are building equity into the wrong container.
The related discipline: use the ™ symbol until a mark actually issues, never ®. Using ® on a pending or unregistered mark is a misrepresentation, and in the US it can prejudice enforcement later. Our own house rule on this business is written down precisely because it is easy to get wrong on a product label six months before registration completes.
Fix the entity question before spending anything on design. A brand is a legal asset that happens to have a visual identity, not the other way around.
The second asset: a website that is the brand's canonical home
A marketplace listing is a page about your product on someone else's domain. A website is the only place on the internet where you decide what your company is, and it is the address every other signal points back to.
This is not sentiment. It is how machines resolve identity.
Search engines and AI answer systems do not really index pages any more — they resolve entities: a company, its products, its location, its people, its category, and the confidence that all of those refer to the same thing. That confidence is assembled from consistency across many sources, and it needs a canonical home to point at. A marketplace seller has no canonical home. Their identity is a storefront ID inside a much larger entity called Amazon, and every citation, every review and every link accrues to that parent, not to them.
The practical version of owning your identity is unglamorous:
- One declared entity. One legal name, one address, one canonical domain, stated identically everywhere. Multiple entities described loosely is the most common self-inflicted wound we see — it fragments the signal.
- Organization structured data on the site, declaring the name, the URL, the logo, the contact route and the profiles that belong to the same company.
- Consistency across every profile that mentions the business. Divergent names and addresses across directories do not add up to more presence, they add up to less certainty.
We rebuilt our own identity on exactly this principle: one declared entity, plain Organization markup, no invented locations. The branding work and the technical work are the same work at this layer.
Google Maps — the honest version
A Business Profile is genuinely valuable. It is a verified, Google-owned statement that your company is real, tied to a place, with a name and a category and a set of reviews that belong to you rather than to a marketplace.
But here is where most advice becomes irresponsible, so we will be direct: a registered-agent mailbox or a virtual office is not an eligible location. Ecommerce sellers routinely try to verify one anyway, because their LLC address is a suite number in a building they have never visited. That is a policy violation, and a suspended profile is a materially worse outcome than no profile — it is a negative signal attached to your name that is tedious to unwind.
The honest decision tree is short. If you have a real premises where you work — a warehouse, a workshop, an office — claim it and treat the profile seriously. If you are a service-area business, register it that way. And if you genuinely have neither, do not fake a location. Put that effort into entity consistency and owned content instead, which is where a product brand's discoverability actually comes from.
Why this is what gets cited
The reason to care about all of this in 2026 is that a growing share of buying research never reaches a blue link. Someone asks an assistant which adhesive-lined tubing to use for a marine splice, and the answer names products and sometimes companies.
To be named, you have to be resolvable. A system generating an answer needs to know that a company exists, what category it belongs to, and that a claim about it can be attributed to a source. A brand with a canonical site, coherent structured identity and substantive content about its own category is legible to that process. A marketplace listing is not — it is attributable to the marketplace, and the marketplace gets the citation.
This is the same asset doing all the work at each stage. The entity makes you real. The website makes you addressable. The consistency makes you resolvable. And a resolvable brand is what gets recommended, which is what produces the visitor who arrives already decided — the visitor who converts at 17% instead of 0.21%.
Which is where the argument closes on itself. Branding is not upstream of sales in some vague, delayed way. It is the mechanism that changes the conversion rate of everything else you do. The organic search work is not a separate line item from the branding work; it is the branding work becoming machine-readable.
The honest limit
None of this is free, and it is genuinely wrong for some businesses.
If you are arbitraging a product with no durable advantage, brand-building is a poor investment — you are constructing identity around something replaceable, and you would do better reinvesting in sourcing and turns. If your margins cannot fund a two-year horizon, do not start: the entity work pays back slowly and the content work pays back more slowly still. And if you are pre-product-market-fit, branding will not rescue you. It amplifies a working offer; it does not create one.
We also have to be honest about our own data. The seller described here has a high-converting channel that is small — brand-shaped demand converts extraordinarily well and there is not yet much of it. Excellent conversion on thin volume is a promising signal, not a finished business. Anyone showing you a conversion rate without the session count is selling you something.
Five questions
- Who legally owns your trademark — the company, or a person? If it is a person, fix that before anything else.
- If your marketplace account were suspended tomorrow, what would you still have? The honest answer is your inventory and whatever you own outright.
- Does your website present you as a manufacturer, or as a reseller of your own product? Most seller-run sites accidentally do the second.
- Is your business described identically everywhere it appears? Every divergence costs you certainty.
- Do you have a real premises? If yes, claim it properly. If no, do not invent one.
If you are profitable on a marketplace and want to know whether the brand transition is worth it for your specific numbers, our client work is mostly businesses at exactly this crossover point. We will tell you if the answer is no.