Skip to content
NameForge

Brainstorm names that will still be standing after the search

Most naming processes run in the wrong order. A founder spends two weeks generating names, falls for one, then searches it and finds a conflict. By then the team has an attachment that makes them argue with the search result instead of accepting it. The fix is to build clearance into the brainstorm, so that only names that have already passed the cheap checks get the chance to be loved.

The process in one line

Frame the constraints, generate far more names than you need, knock out with free searches in batches, shortlist the survivors, and only then test and choose. Pay a professional for a full search on the final two or three, not the first twenty.

Step 1: frame it before you generate anything

Write down, before any names, the facts that decide whether a name can work:

Step 2: generate wide, and on purpose

Aim for 50 to 100 candidates. Most will not survive, and that is the point. Generate deliberately across the distinctiveness spectrum, leaning towards the protectable end:

Name types to generate, roughly from most to least work to establish, and from strongest to weakest protection.
TypeHow to generateClearance outlook
InventedNew words from roots, sounds or blendsUsually the cleanest. Check for meanings in your other markets' languages
ArbitraryReal words from unrelated fields, such as nature, places, objects or historyGood, if no one in your trade already uses the word
SuggestiveMetaphors and associations that hint at the benefit without stating itMixed. Popular metaphors in a sector get crowded
DescriptivePlain statements of what you doPoor. Hard to register, hard to enforce, and usually already in use

Name generators, including AI tools, are useful for volume and for getting past the first obvious ideas. Treat their output as raw material. A generator does not check registers, and it draws on the same common patterns that many other businesses have already used, which is why its suggestions often turn out to be taken.

Step 3: knock out in batches with free searches

A knockout search looks for obvious conflicts. It is not clearance. It removes the names that are clearly unavailable so you do not waste time on them. Run every candidate through:

  1. An exact-match trademark search in your markets: the USPTO's trademark search, the UK IPO search, TMview for the EU and national registers, and WIPO's Global Brand Database for international registrations. Look at live marks in your classes and in related ones.
  2. A similarity glance on anything close. Search the core syllable and obvious spelling variants, not only the exact word. Why identical is the wrong test explains why.
  3. The company register of your incorporation country. A near-identical name will simply be refused.
  4. A web search with the name and your industry terms. Businesses that trade without a registration still have rights, and a search engine is the cheapest way to find them.
  5. A domain lookup on the endings you actually need. The checker runs live registry lookups.

Cut ruthlessly. A name with a live identical mark in your class goes. A name with a possible conflict goes unless it is far better than the alternatives. You will have fewer names to argue about, which is useful in itself.

Step 4: test the survivors

Put the 5 to 10 survivors through the human tests that registers cannot do:

Step 5: the professional search, on two or three names

Once you have a first choice and a back-up, pay for a full clearance search and an opinion in the markets that matter. A professional search covers similar marks, common-law use and the registers a free search misses, and the opinion assesses risk in a way a database cannot. Paying for this on two or three names is affordable for most new businesses. Paying for it on twenty is not, which is why the earlier steps exist.

Then move quickly. Register the domain, file the trademark application and incorporate in the order set out in the pre-commit checklist, before you announce the name publicly.

Questions

How many candidates normally survive the knockout stage?

It depends heavily on the sector and the name types. Descriptive names in crowded fields such as software or consulting are eliminated at high rates. Invented names survive more often. If almost nothing survives, the problem is usually the vocabulary you are generating from, not bad luck.

Is a free knockout search enough if I am a small local business?

It lowers the risk considerably and is far better than nothing. It cannot assess similarity the way an examiner or court would, and it misses unregistered use that does not show up online. How much more to spend depends on how much the name will carry: signage, packaging, a website and years of reputation all raise the cost of being wrong.

Should we vote on names as a team?

Vote only on names that have passed the knockout stage. Voting earlier creates attachment to names that may not survive the search. When one of them fails, the team's favourite has just been rejected, which is a worse position than never having picked it.

Related

Published 2026-10-18. A process guide. It does not replace a professional clearance search or an opinion on any specific name.