Skip to content
Domain Names

How to choose between two names you both like

9 min read

Two names. Both good. You have been going back and forth for four days and the argument has stopped producing new information.

That is what happens when a decision is made on 'which feels better', because feel is not stable — it moves with mood, with whoever spoke last, and with which one you looked at most recently. The way out is to replace one vague comparison with six specific ones, each of which has an answer.

None of the tests takes long. Run them on both candidates in the same sitting, write the results down, and the choice usually makes itself. Where it does not, the last section covers what to do about that.

Stop asking which one is cooler

When two candidates are close, intuition loops rather than converging. Every time you look at one it feels like the answer, and the feeling reverses within a day.

So turn it into a comparison against explicit criteria. This is not about removing judgement — the weighting at the end is entirely judgement. It is about making sure the judgement is applied to something other than recency.

One rule for every test below: run both candidates through it with the same people, in the same session, without saying which one is yours. Any test where the participants know your preference is measuring your preference.

1. Pronunciation, side by side

Show both names written down to people who have never seen either, and ask them to read each aloud. Record where they hesitate rather than which one they say they like.

A hesitation is a small permanent cost. It happens in every sales call, every introduction and every conference badge, and a candidate that produces hesitations will produce them for as long as the company exists. Two people pronouncing the same name differently is worse still: it means your own customers will disagree about how to say your brand.

2. Spelling, from sound

Now say each name and ask listeners to type what they heard. Count both the number of errors and their kind.

The kind matters more than the count. A single recurring ambiguity — everybody choosing the same wrong vowel — is one predictable problem and sometimes an opportunity, because that spelling may be available and may be the better name. Five different guesses at the same word is an unfixable problem.

This test regularly outweighs style preference, because it measures lost direct traffic and word-of-mouth friction rather than taste. The underlying mechanics are in what makes an invented name easy to say and spell.

3. Delayed recall — the test that predicts most

Show both candidates among four or five distractor names, talk about something else for ten minutes, then ask which names the person remembers. Do not reveal which two are yours and do not lead the question.

Immediate preference measures which name is cleverest. Recall measures which name is retrievable, and retrievable is what you are actually buying — a customer who cannot reconstruct your name a day later cannot recommend you, search for you, or type your address.

What makes a name memorable, and what does not

Worth knowing before you run the test, because the results are otherwise mysterious. Memorability is not shortness. A short string with no rhythm, no recognisable structure and no obvious pronunciation is forgettable at four letters. What memory needs is the ability to encode a name and then reconstruct it, and there are four things that help.

  • Clear pronunciation, which gives the name a sound as well as a shape — two routes to recall instead of one.
  • A clean rhythm. Names with an obvious beat get repeated accurately in conversation; names whose stress is ambiguous get repeated approximately.
  • A hook: a real word, a metaphor, or a recognisable fragment the mind can attach the name to.
  • Recoverable spelling. Remembering the sound is useless if the letters cannot be reconstructed from it.

And one thing that hurts: too much distinctiveness. One unusual feature aids recall by giving the name a handle. Three unexpected spellings turn it into a puzzle, and a puzzle is not remembered, it is abandoned.

Note that the extension is part of the memory task. Somebody has to recall the name and where it lives, and every character of that is something to carry.

4. Domain friction

The cleaner brand is not always the cleaner address, and this is where the two things come apart. Compare the full domains rather than the names: the exact match, the extension, any modifier one of them needs, and whether either could be confused with a neighbouring domain.

It is entirely normal for the better-sounding name to be the one that needs 'get' in front of it, and for the slightly less exciting name to have a clean exact match. That trade is real and it should be made deliberately.

Does the domain have to match the business name?

No, and it is worth being clear about why exact matches are attractive so you can price the alternative. An exact match removes explanation: somebody who hears the brand can guess the address, and the email addresses reinforce the same spelling every time you send one.

A short modifier is a reasonable substitute when it sounds natural said out loud and does not change what the brand is. Where it stops being reasonable is when the address looks like a different company — then customers have two names to learn instead of one, and that cost grows with every word-of-mouth referral you get.

How much this matters depends on how people find you. If most arrive by clicking a link, exact matching matters less. If they arrive because somebody told them your name in a conversation, guessability is close to everything.

One firm rule regardless: pick a single canonical address and use it everywhere, redirecting anything else to it. Two live addresses for the same company means customers guessing which one is official.

5. Distinctiveness in search

Search both names, and their near-spellings, and look at what territory each one lands in. One candidate may sit in a crowded field where you will compete for your own name for years; the other may be clear.

This is not primarily an SEO question — it is a 'can a customer find you' question. A name you share with an established company in an adjacent market means every recommendation has a chance of arriving somewhere else.

6. Future flexibility and fit

Imagine each name attached to the business two years from now, with a second product, a second country and a larger audience. Which one survives that with less explanation?

Then do the fit test properly rather than by instinct: write down three adjectives the brand should communicate, before scoring anything, and rate both candidates against those three. Deciding the adjectives first is what stops the exercise becoming a way of justifying the name you already wanted.

When it is a great brand against an available domain

This is the version of the choice that causes the most agonising, and it is usually posed too narrowly: the perfect brand with no exact .com, against a mediocre brand that is available. Framed that way it has no good answer.

The framing is wrong in two directions. A weak name does not become strong because its .com is free — pronunciation, memory and distinctiveness will affect the business every single day, and availability will affect it once. But a strong brand with a confusing address does give away part of its advantage, in every referral and every spoken introduction.

So do three things before choosing between them. Check what is actually on the exact match, because an active company in your market makes this a brand conflict rather than a domain problem. Try a clean modifier and read the whole address aloud. And then generate fifty to a hundred more candidates around the qualities you liked in the strong name — the most common outcome is a third option that beats both, and comparing one perfect name against one bad available name is a false choice you constructed.

Stage matters too. An experiment can absorb more domain compromise than an established business with real direct traffic, because the cost of changing later scales with everything you built in the meantime.

Score brand strength and domain quality as two separate numbers rather than one impression. That is what stops a single emotional variable — usually the satisfaction of owning the exact match — deciding the whole thing. What you are looking for is the strongest total system: brand, address, distinctiveness and room to grow, with the least damaging trade-off. Where the domain side is close, .com vs .io vs .ai vs .co has the costs.

Research before you declare a winner

A candidate can win every test above and still be the wrong choice. Before committing, search both for existing companies, products and registered marks in the markets you will operate in.

It is worth doing this on the loser as well as the winner. If the research kills your first choice, you want to already know whether the runner-up is clear — from shortlist to registered covers what to search.

Weight the criteria, then actually choose

Not every test matters equally, and the weighting is where your business enters the decision. A podcast weights spoken clarity above almost everything. Enterprise software weights email clarity and credibility. A consumer app weights recall and how the name looks in an icon.

The comparison, on one line each

  • Read aloud from the spelling — where did people hesitate?
  • Spelled from sound — how many errors, and were they the same error?
  • Remembered after ten minutes of distraction?
  • Full address: exact match, extension, modifier, confusable neighbours
  • Search territory: crowded or clear?
  • Still works with a second product and a second country?
  • Scored against three adjectives chosen in advance
  • No conflicting company or mark in your markets

If the weighted result is close, accept that there is probably no objectively correct answer and choose the one you would rather say out loud a thousand times. A name is built into something after the decision at least as much as during it, and a fortnight spent deciding is a fortnight not spent building the thing that makes the name mean anything.

In short

Replace 'which feels better' with six tests run on both candidates in one sitting: read aloud, spelled from sound, recalled after a delay, address friction, search territory, and room to grow. Then weight them for your business and decide the same day.

Frequently asked questions

How many people do I need for these tests?

Three to five per test, and none of them can know which name is your favourite. You are not measuring popularity; you are looking for whether the same mistake happens twice.

What if both names pass every test?

Then you have two good names and the decision is genuinely low-stakes, which is worth knowing. Pick the one you would rather say a thousand times, register it today, and keep the other on the shortlist.

Does the domain need to match the business name exactly?

No. An exact match saves explanation and a sensible modifier is a reasonable substitute. What causes real damage is an address that reads as a different company, because customers then have two names to remember.

Should I let a vote decide it?

Use tests, not votes. A vote among people who know your preference measures your preference, and a vote among strangers measures immediate appeal — which is the one property that has almost no relationship to how well a name works over years.

Sources

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

Related guides