Naming is a strategy problem

Naming often gets left for the end, as a quick brainstorm once the product is done. But it’s one of the first strategic decisions you make, and one of the hardest to change later. Here’s how we approach it: positioning first, then the kind of name, then a long list, then a hard filter.

Start from positioning, not from words

Before any candidates, answer the strategy questions. What is this for, who is it for, and what does it need to claim in a crowded category. A name is a position compressed into a word. If you do not know the position, you are not naming, you are guessing, and a clever word attached to an unclear strategy is just a nicer way to be forgettable.

The four kinds of name

• 01Descriptive: it says what the thing is, like General Motors or Whole Foods. Instantly clear, hard to own, and it can box you in when the company outgrows the description.
• 02Suggestive: it evokes a quality without stating it, like Stripe or Vercel. A strong middle ground, clear enough to hint and open enough to grow.
• 03Abstract: a real word borrowed from elsewhere, like Apple or Amazon. Little inherent meaning, which is the point: you get to fill it with your own.
• 04Coined: an invented word, like Google or Spotify. Maximum ownership and distinctiveness, but you pay to teach the world what it means.

Naming a SaaS or AI product

• 01Avoid “AI” in the name unless AI is the product forever.
• 02Avoid naming after one feature you’ll outgrow.
• 03Say it on a sales call: can someone spell it after hearing it once?
• 04Domain: an exact .com is nice, not required. A clear prefix (get-, use-, try-) or a strong .ai or .io is fine early on.

A live example: our own name

FormForm is a coined, repeated name: easy to own, but people search for it as ‘form form form’ or ‘form in form’. Coined names trade instant clarity for ownership, and you pay for it in spelling. We decided that was worth it.

Generate wide, then filter hard

Once you know the position and the kind of name, generate without mercy and judge without sentiment. The generating is the cheap part. The filtering is where the strategy actually happens.

• 01Distinctiveness: does it stand apart in the category, or blur into the sea of names that rhyme with the leader.
• 02Durability: will it still fit in five years, or does it pin you to a feature, a year, or a trend you will outgrow.
• 03Room to mean more: can it absorb meaning over time, or does it say everything on day one and nothing after.
• 04Sound and rhythm: say it out loud, in a sentence, on a phone call. Names live in the mouth before they live on the page.
• 05The boring, fatal checks: trademark, domain, and what it means in the other languages your customers speak.

Where AI helps and where it stops

Generation is the half that just got free. Ask a model for five hundred candidates in the suggestive register and you will have them in seconds. That is genuinely useful, and it is not the job. The job is deciding which one is worth defending for a decade, which fight is worth picking, which compromise you can live with. That is judgment, and it is exactly the part a generator cannot supply. Use the machine for breadth. Reserve yourself for the choice.

Naming direction is part of our brand identity work, and part of a Launch sprint.

• Strategy
• Identity
• Process