How do you name a product?
A good product name is one that survives contact with reality: it's easy to say, hard to confuse, and doesn't box you in when your product evolves. The goal isn't to be clever — it's to be remembered and trusted. Get the name directionally right, then ship; you can always rebrand, but you can't recover lost time.
Start with what the name has to do, not what it should sound like
Before you brainstorm a single word, write down the job the name must perform. Does it need to explain the product instantly (useful for B2B tools with short sales windows)? Or does it need to anchor an emotional identity (useful for consumer brands where trust is built slowly)? These are different briefs. A name like 'Checkr' tells you almost nothing about background checks on first hearing but works because the company had the budget and time to teach the market. A name like 'Zoom' worked partly because it felt fast. Know which game you're playing before you start generating options.
The most common mistake founders make is optimizing for what sounds cool in a Slack thread rather than what works in a customer's mouth. Test your shortlist out loud. Say it to someone on the phone and ask them to spell it back. If they get it wrong, that's a real cost — every misspelling is a lost direct-navigation visit, a confused referral, a support ticket about the wrong account. Phonetic clarity is not a nice-to-have; it's a functional requirement.
Also think about the name's ceiling. Founders routinely pick names that are too narrow — a city name when they plan to expand nationally, a specific use-case name when the product will generalize. Amazon didn't call itself 'OnlineBooks.' Leave yourself room to grow without making the name so abstract it means nothing.
The mechanics: how to actually generate and filter candidates
Run a structured generation session, not an open-ended brainstorm. Set a timer for 20 minutes and generate at least 50 candidates without filtering. The first 15 will be obvious; the useful ones come after you've exhausted the obvious. Use constraints to force creativity: what's the shortest possible name? What's the most literal? What's a metaphor from an adjacent domain (finance, biology, geography)? Constraints produce more actionable output than 'just think of good names.'
Once you have a list, filter in two passes. First pass: eliminate anything that fails the phone test (too hard to spell or say), anything with trademark conflicts in your category (run a USPTO search and a Google search for 'sitename.com review'), and anything that has an embarrassing meaning in a language relevant to your target market. Second pass: rank what's left by memorability. Read each name once to five people who don't know your product and ask them to recall it an hour later. The ones they remember are the ones that work.
Don't wait for the perfect name. Founders spend weeks on this and ship months late. A decent name shipped early beats a perfect name shipped never. The history of successful companies is full of names that seemed bad at the time — 'Apple' was thought to be too casual for a computer company, 'Slack' was thought to have negative connotations. Execution eclipses naming.
Domain, trademark, and handle: the three things you must check
Once you have 3–5 finalists, run them through a hard checklist before you commit. First, the domain: you want the exact .com if you're building a consumer product. For B2B, a short qualifier like 'get' or 'use' as a prefix is acceptable (getdropbox.com was Dropbox's early domain). Avoid hyphens — they're invisible in verbal communication and look cheap. If the .com is owned by a squatter, make an offer before you build the brand; buying a domain for $5,000 after you've raised a round is much cheaper than rebranding after you've printed 10,000 business cards.
Second, trademark: search the USPTO TESS database yourself before paying a lawyer, then pay a lawyer to do a proper clearance search before you file or commit. A name that's already registered in your category is a ticking clock — you may be able to launch, but you'll face a cease-and-desist the moment you have any visibility. This has killed companies. Budget $1,500–$3,000 for a proper clearance search and filing. It's the cheapest insurance you'll ever buy.
Third, social handles: check Twitter/X, LinkedIn, Instagram, and TikTok. The handle doesn't have to be a perfect match — a slight variation is fine — but you want to avoid a situation where a large, active, unrelated account owns your exact name on the platform where your customers spend time. A brand collision on social is a slow-burn customer confusion problem that compounds over time.
When to rename and when to live with what you have
Most founders think about renaming too early (when the name is just mildly awkward) or too late (when they're finally embarrassed enough to act). The right time to rename is before significant brand equity has accumulated — which in practice means before you have more than a few thousand customers who know you by name, before you've done significant press, and before your domain has meaningful SEO authority. After those thresholds, the switching cost grows quickly.
If you're post-launch and genuinely stuck with a bad name, the practical move is to treat the rebrand as a product launch. Pick a date, build toward it, and announce it with enough context that customers understand what changed and why. The companies that damage their brands in renames are the ones that do it quietly and confusingly — customers feel like something was hidden from them. Transparency about why you're changing and what stays the same (your mission, your team, your product) is the entire communication job.
Garry Tan's principle of leading with the point and being concrete rather than consultant-speak applies to naming too: the best names say what a thing does or how it makes you feel, not what the founder thought sounded sophisticated in a conference room. When in doubt, bias toward the obvious.
The one thing to do
Pick a name that passes the phone test, clear it with a trademark attorney, secure the domain, and ship — a good-enough name launched fast beats a perfect name that delayed you six months.
Frequently asked questions
Does the name matter that much if the product is great?
A great product can overcome a mediocre name, but a bad name creates real friction: misspellings, misdirected traffic, confused referrals, and trademark risk. It's not the most important thing, but it's not free to ignore either.
Should I get a .com domain or is another TLD okay?
For consumer products, .com is strongly preferred — users default to it and you'll lose direct-navigation traffic to whoever owns the .com. For B2B SaaS, alternatives like .io or .co are more accepted, but the .com is still worth buying even if you don't use it as your primary domain.
How many people should I get feedback from before deciding?
Five to ten people is enough for the recall test; beyond that you're getting noise, not signal. What matters is that they're representative of your actual customers, not your co-founders' friends who will be supportive regardless.
What if the name I want is trademarked in an unrelated industry?
Trademark protection is category-specific, so a name registered for agricultural equipment may be perfectly usable for software. Get a lawyer to confirm the specific classification overlap — this is not a call to make yourself.
Sources
- gstack: skillify/SKILL.md — Garry Tan