Skip to content
Domain Names

The complete domain naming checklist, idea to registration

8 min read

Searching for a domain should not feel like pulling a slot-machine lever: type a name, see 'taken', sigh, repeat. That loop produces bad names, because the only thing it selects for is what happens to be unregistered.

A structured process gives you more good candidates, wastes fewer checks and makes the final decision less emotional. This is the whole sequence, in order, with the two questions people always ask — how many ideas, and when do I stop — answered where they arise.

It is designed to be read once and then used as a list. Nothing here takes long; the value is entirely in doing the steps in this order rather than a different one.

Why the order matters more than the effort

Naming has two activities that want opposite mindsets. Generating wants abundance and no judgement. Choosing wants strictness and no attachment. Almost every naming disaster comes from doing them simultaneously.

The fix is a funnel: generate broadly, filter hard, check in batches, research the finalists, then decide. Availability arrives late, as one filter among several, rather than early where it quietly becomes the only one.

This is also what makes any tool useful. Software can produce hundreds of strings; only your process can decide which strings deserve attention. Filtering for quality before checking availability is what turns volume into choice rather than noise.

Steps 1–3: decide what you are looking for

Skipping this is why most naming sessions drift. Twenty minutes here saves the whole afternoon.

1. Write the naming brief

Audience, product, personality, geographic scope, a comfortable length, whether it must work in languages other than English, whether you insist on .com, and how far the company might travel from what it does today. Half a page. These constraints become the criteria you score against at the end, which is how you avoid a decision made purely on whose turn it is to be stubborn.

2. Identify naming territories

Collect directions rather than names: customer outcomes, emotions, metaphors, category concepts, and pure sounds you like. Four or five distinct territories beats one keyword searched exhaustively — and if a whole territory turns out to be unavailable, you have others rather than nothing.

3. Choose seed words

Pick five to ten roots that represent those territories. Seeds are what give generated variations coherence: candidates descended from a seed share a personality, while candidates from an open-ended request share nothing at all.

Step 4: generate broadly — and how many is enough

There is no correct number, and the honest answer depends on how constrained you are. A local business with a descriptive name needs far fewer candidates than a startup insisting on a short .com.

Ten is almost always too early. With a set that small you are choosing the best of a weak field rather than discovering a strong direction, and you have no way to notice which patterns are working.

Several hundred raw ideas is entirely reasonable, because most raw ideas are supposed to fail. A large first pass gives you permission to delete without feeling that every candidate must be rescued — and that permission is the actual point of generating a lot.

Work in rounds rather than one endless list. Thirty to fifty around one seed, look at what came back, then another round built on the strongest fragment. Each round is better informed than the last. A single list of three hundred names teaches you nothing between the first and the three-hundredth.

Steps 5–7: filter on speech, spelling and memory

Three passes, in this order, before any domain is checked. Each one is a different failure mode and a different test.

5. Filter for pronunciation

Show candidates to people cold and ask them to read them aloud. Remove anything that repeatedly causes a hesitation, unless the unusual pronunciation buys something exceptional — and be sceptical, because it rarely does.

6. Filter for spelling

Now reverse it: say each candidate and ask people to write down what they heard. Note where errors repeat and where two spellings are equally plausible. A name that arrives in four different spellings will send a quarter of your direct traffic somewhere else. The mechanics behind both passes are in what makes an invented name easy to say and spell.

7. Test memorability

Show four or five, change the subject for ten minutes, then ask which are remembered. Delayed recall is a much better predictor than asking which one people prefer — preference measures cleverness, recall measures stickiness, and you are buying the second one.

Step 8: check domains in batches

Now, and not before. Check the serious candidates together so results are comparable, and compare the exact .com against sensible alternatives and extension trade-offs in the same sitting.

Two things to hold onto here. Availability is not permission to redefine quality — a name that failed step 5 does not become good because it is free. And a result is a snapshot rather than a reservation: what a domain checker actually knows covers why one tool says free and another says taken.

What the batch tells you also decides the next round. If an entire naming family comes back unavailable, choose deliberately between modifying that family and moving to a different seed. What you must not do is start bolting increasingly awkward characters onto names that were taken for a reason.

Steps 9–11: research what survives

This stage removes candidates that passed every creative test, and that is normal — it is precisely why you generated alternatives.

9. Research existing uses

Search the exact name and near-spellings across the web, company registers, products, app stores, social platforms and the markets you intend to sell in. You are looking for anything that would make a customer repeatedly distinguish you from somebody else.

10. Research trademarks

Use the official databases for the jurisdictions that matter — the USPTO for the United States, EUIPO for the EU, WIPO's Global Brand Database for a broad first look. Domain availability is not legal clearance, and a high-stakes launch in a crowded category warrants professional advice rather than a confident afternoon.

11. Inspect history and economics

For any domain that was previously owned, look at what was published there. For every domain, check the renewal price rather than the first-year price, and read the registrar's transfer policy before you are trying to leave.

Steps 12–13: test in context, then score

By now you should have three to ten serious contenders. Not three hundred unranked names in a spreadsheet — a working shortlist has to be small enough to compare carefully, and anything larger is a way of avoiding the comparison.

Put each finalist where it will actually live: a homepage headline, an email address, an app icon, a browser tab, a sentence in a sales call, a podcast mention, a support conversation. Names fail in specific places, and the places are cheap to check.

Then score them on the same criteria, and weight the criteria for your business rather than in general. A podcast weights spoken clarity heavily. Enterprise software weights email clarity and credibility. A consumer app weights the icon. The weighting is the decision; the scoring just makes it visible.

Score every finalist on the same axes

  • Easy to pronounce from the spelling
  • Easy to spell from the sound
  • Remembered after a delay
  • Distinctive in your category
  • Right for the audience in the brief
  • Room to grow past the first product
  • Clean domain strategy — no modifier you have to explain
  • Sensible acquisition and renewal cost
  • No obvious market confusion
  • Conflict and trademark research completed

Knowing when to stop generating

The most common failure at this point is not a bad choice, it is no choice. More generation always feels productive and eventually is not.

The signal is clean: if two consecutive rounds produce nothing that scores above your current top three, you are procrastinating rather than exploring. Set a decision date when you build the shortlist. A shortlist is only valuable if it becomes a choice.

Steps 14–15: register, then build

Register the winner the same day you decide. Use a registrar you would be happy to still be with in five years, turn on auto-renew and two-factor authentication, put a contact address on it that does not live at the domain itself, and keep the renewal date somewhere you will see it. From shortlist to registered has the full checkout checklist.

Then stop thinking about the name. Registration is the start of the work, not the end of it: what eventually makes the domain valuable is the product, the consistency and the customer experience behind it. Almost every name people admire was unremarkable on the day it was chosen.

In short

Brief, territories, seeds, generate broadly, filter for speech and spelling and memory, then check availability — in that order. Availability arriving late is the difference between choosing a name and being handed one.

Frequently asked questions

How long should naming take?

For most projects, days rather than weeks. The generating and filtering is an afternoon; the research is another; the gap between them is for letting the finalists sit long enough that the exciting one stops being exciting.

Should I check every name I generate?

No. Check only what has already passed your pronunciation and spelling filters. Checking everything is how the results of a checker end up choosing your brand.

Is an available .com automatically a finalist?

No, and treating it as one is the most expensive habit in this whole process. Availability qualifies a candidate for consideration; it does not tell you anything about whether the name is good.

Do I need the exact .com?

Not necessarily, and availability establishes no rights in the name either way. What the exact match buys you is people typing your address correctly on the first attempt, which is worth a lot and is not worth everything.

If I had to keep one test, which one?

Say the name once to somebody who has not seen it, and ask them to type it. It is the cheapest test and it predicts more of the day-to-day cost of a name than everything else combined.

Sources

These support factual background only. The guide itself is original writing, and nothing here should be read as legal advice.

Related guides