Naming a Product vs Naming a Company

The two poles, in founder terms

Every naming architecture sits between two extremes. At one end is the branded house: one master brand covers everything, and new products arrive as descriptors rather than new words. At the other is the house of brands: each product carries a standalone name and the parent stays invisible, so a customer may never learn that two things on the same shelf share an owner. The branded house concentrates every scrap of attention into one word, which is why a company with no marketing budget gets more from it than any other structure; the house of brands spends attention across many words to buy freedom, the room to sell to incompatible audiences at incompatible prices.

Between the poles sit two genuinely distinct models, sub-brands and endorsed brands, from the same received taxonomy (David Aaker and Erich Joachimsthaler set it out as a spectrum). Naming them matters, because “something in between” is not an architecture and does not survive its first disagreement.

The models, and what each one costs

Brand architecture for startups is usually taught with portfolio examples a founder cannot act on. Here is each model, with what it buys and what it charges, in order of how much extra naming work it takes on.

In practice most companies end up hybrids, a branded house for the core with one endorsed or standalone name off to the side. That is fine as long as it is a decision. The expensive version is the hybrid nobody chose, where a codename, an acquisition, and a project name all reach customers because no one ever ruled on it.

What every extra name costs you, forever

This is the part of the decision that rarely gets said out loud. A name is not an asset you buy once. It is a subscription, and adding a second name doubles a set of costs that never stop being paid:

1. A second slot in someone else’s memoryRecognition is not transferable. A customer who knows your company by name does not therefore know your product by name, and the second word has to be repeated into place from zero, using the same finite attention that was building the first one.
2. A second domain positionA second name wants its own domain, its own defensive variants, and its own renewals. Whatever you find is worth checking directly at the registrar rather than treating as settled, because registration data moves by the minute.
3. A second cross-language readEvery market you enter has to be screened again for the new word. A screen that reports no flags in the languages checked is a real result for one name only, and says nothing at all about the other.
4. A second trademark positionA separate name is a separate mark, in separate classes, in every jurisdiction you care about, with its own searches, its own filings, and its own renewals. This is general orientation and not legal advice, and it is the cost founders underestimate most.
5. A sentence of explanation, foreverOn every sales call, in every deck, in every piece of copy, someone has to say how the two names relate. It takes six seconds, it never stops, and it repeats across every conversation your company will ever have. That is the real price of the second name.

None of these is dramatic on its own, which is why they get waved through: the founder is thinking about a logo and a landing page, not five recurring bills. And the asymmetry is the trap. The benefits of a second name are speculative and arrive later; the costs are certain and start immediately. That is not an argument for never doing it, but for making the case explicitly, in writing, before you do.

The four questions that decide it

Should I name my product after my company? Answer these four questions about your actual situation, today, and write the answers down. Each one is a yes or a no, and vague answers count as no.

1. Is the buyer the same person?Not the same market, the same human being making the decision. If the person who chooses this product is the person who already knows your company, a second name asks them to learn a word they did not need. If it is a different buyer, in a different room, with different reasons, the separate name starts earning its keep.
2. Does the company name foreclose this?Say the company name out loud next to the new product and listen for the flinch. A company named for a category, a technology, or a region will actively argue against a product that sits outside it. If the name does no arguing, you do not need a new one.
3. Would you ever sell or spin this out on its own?A product you might one day detach needs a name that can leave. Naming it after the company welds it to the company, and unwelding it later means a rename at the exact moment you are trying to close a deal. If the honest answer is no, this question costs you nothing.
4. Can you fund the second name for three years?Not buy it, fund it: the repetition, the domain, the language screens, the mark, and the explaining. If nobody on the team has the time to teach the market a second word while still teaching it the first, you cannot afford it yet, and the honest move is to wait rather than to launch a name you will neglect.

Now count the yes answers. Zero means a branded house: name the product after the company and put the saved effort into making one name known. One means a sub-brand or an endorsed brand, depending on how far the product needs to stand apart. Keep the parent attached in every appearance, and treat that attachment as a rule rather than a habit. Two or more means the product has earned a standalone name, and you should budget for it as a permanent line item rather than a one-off.

The fourth question can veto the others: a product that deserves its own name but has nobody to feed it is better served by a descriptor for now. One case the questions do not name directly is price: a cheap line under a premium name can drag the premium name down, which is exactly what the endorsed model exists to prevent, so a very different price point counts as a yes on its own. Write the answers down, because architecture arguments recur, and a dated page of four answers settles the same debate every time it comes back.

Note what is not on the list: “the team is bored of the name”, “it feels like its own thing”, and “an investor suggested it” do not count. This decides how many names you need; what kind of word each should be is a separate question, in descriptive vs. abstract brand names. If your real question is whether the existing name is wrong rather than whether to add one, that is when to rename your startup.

The internal codename that becomes the real name

Teams give projects placeholder names because work needs a handle before it needs a brand. The trap is that a codename leaks in a support ticket, a status page, a release note, or a URL, and by the time anyone asks whether it is a good name, customers are already using it and the answer has stopped mattering.

The best documented case is Bluetooth. Jim Kardach of Intel has written that he proposed it in 1997 as a codename, after the tenth-century Danish king Harald Bluetooth, expecting marketing to replace it. The intended replacements were RadioWire and PAN; PAN was already heavily used, the RadioWire trademark search could not finish before launch, and by then the codename had spread through the industry. The placeholder became the name of a global standard. Note the survivor bias, though: the codenames that leaked and did not work out are not written up anywhere.

There are two defensible policies. Either treat every codename as a real candidate, so it passes the same checks as anything else before an outsider sees it, or make codenames deliberately unusable: unpronounceable, or plainly silly enough that nobody could ship them by accident. The failure mode is the middle, a placeholder just good enough to survive, which is how a company acquires a second permanent name without ever deciding to.

What happens when one product wins

If you do carry two names, one common ending is that the market picks for you. A single product becomes the reason anyone knows you, customers start using the product name for the whole company, and the company name turns into paperwork nobody says out loud.

Two well-documented cases show both ways this resolves. Tiny Speck built an internal chat tool while making a game called Glitch; the game shut down at the end of 2012, the tool became the business, and the company renamed itself Slack Technologies in 2014. 37signals went the same way and then came back: it renamed itself Basecamp in 2014 to focus on that single product, and reverted to 37signals in 2022 once a one-product name no longer fitted a company shipping more than one thing. A name that perfectly describes the company’s current shape is a bet on that shape holding, and letting the winning product name absorb the company is easy to do and expensive to undo. If your company name is already being ignored in favour of a product name, that is a signal worth acting on rather than drifting through.

And if you find yourself weighing two specific candidates for whichever name you decide you need, score them against each other rather than arguing: how to choose between two business names.

Frequently asked questions

Should my product have a different name than my company?

Usually not at the start. One name means every mention builds one memory, one domain position, one language screen, and one trademark position instead of two. Buy a second name when it does a job the company name cannot: a different buyer, a different price point, something you may sell separately, or a category the company name shuts out.

What is the difference between a branded house and a house of brands?

In a branded house one master brand covers everything and new products arrive as descriptors, as in FedEx Express and FedEx Ground. In a house of brands each product carries a standalone name and the parent stays invisible to customers, the way Procter and Gamble sells Tide, Pampers, and Gillette. Sub-brands and endorsed brands are the two genuinely distinct middle models.

What is brand architecture for a startup?

It is the decision about how many names you carry and what each one has to do: the company, the products, and anything the company signs its name to. For most startups the correct architecture is one name, revisited when a second name has a specific job. Architecture is a separate question from name style, which is about what kind of word to pick.

Can I name my company after my product later?

Yes, and companies do it in both directions. Tiny Speck renamed itself Slack Technologies in 2014 after its internal chat tool outgrew the game it was built for. 37signals renamed itself Basecamp in 2014 to focus on one product, then reverted to 37signals in 2022 once it shipped more than one. Later is possible; it is simply more expensive than choosing deliberately now.

Does Namoly score a product name and a company name together?

No. Namoly scores one name at a time on four deterministic checks: domain availability, memorability and distinctiveness, phonetics and readability, and cross-language safety. If you are carrying two names, run each one separately and read them side by side. It checks neither trademarks nor social handles.