Brand Name Ranker

What this generator does

Ranks your shortlist from your own answers to a series of two-way questions, using a merge sort in which every comparison is a question put to you. That asks far fewer questions than comparing every pair: about 10 for six names against 15, and about 23 for ten against 45. The order it produces is exactly the one your answers imply. It also offers to test whether those answers were consistent with each other, by asking about pairs it never needed to ask about — which is the one thing this method cannot guarantee on its own.

How to use this tool

1. Paste your shortlisted names, one per line.
2. Press Start comparing.
3. For each pair, pick the one you prefer. Undo if you misclick.
4. Read the ranking, then use Check my answers to see whether your preferences were consistent.

Understanding the controls

Your shortlisted names

One per line, in any order — where a name starts in the list has no effect on where it finishes. Trim to real contenders first: ten names you care about give a better result in half the questions than thirty with padding.

Which do you prefer

The only question asked. There is no scoring and no ties, which is what keeps the number of questions down. Judge each pair on the same basis — switching between 'sounds better' and 'is more available' partway through produces an order that means nothing.

Check my answers

Asks about pairs the sort skipped because your other answers already implied them. Agreement means your preferences form a consistent order. Disagreement is common, and means the shortlist is not as settled as the ranking makes it look.

Worked examples

Six names

About 10 questions instead of 15.

Ten names

About 23 questions instead of 45.

Twenty names

About 64 questions instead of 190.

An inconsistent shortlist

The check finds you prefer C to A, while the ranking puts A first.

Common use cases

• Settling on one name from a shortlist you have been staring at for a week
• Getting a team to agree an order without the loudest voice deciding it
• Ranking product or feature names before taking them to a stakeholder
• Working out whether you actually have a favourite or are genuinely torn
• Producing a defensible order, with the number of comparisons behind it

How this generator works

The list is sorted by merge sort, with the comparisons replaced by questions to you. Merge sort splits the list down to single names and then merges sorted runs in pairs, and merging two sorted runs only ever needs to compare their leading items — which is why most pairs never have to be asked about. Once you have said A beats B and B beats C, no question about A and C is put, because the answer already follows. Measured over 200 random starting orders, ten names take 22.6 questions on average and twenty take 64.3, against 45 and 190 for every pair; the theoretical minimum for those sizes is about 22 and 62, so very little is being wasted. The finished ranking is read straight off the single remaining run.

Randomness and fairness

Nothing about the ranking is random. The questions follow deterministically from your list and your answers, and the same answers always give the same order. Randomness appears only in choosing which unasked pairs to offer as consistency checks, and that does not affect the ranking.

For how randomness is produced across the whole site, see how Generate Random works.

Assumptions this tool makes

• Every question is judged on the same criterion. No tool can detect a change of basis partway through.

Limitations and good to know

• It ranks preference, and preference is not a business case. It cannot tell you that a name is unavailable, already trademarked, hard to spell over the phone, or rude in another language.
• It assumes your preferences are transitive — that liking A over B and B over C means you like A over C. Where that fails, the order depends partly on which questions came up. The consistency check exists to find that rather than hide it.
• There are no ties. Two names you rate exactly equally will still be separated by whichever you clicked.
• A group answering together produces the group's ranking on the day, which is not the same as everyone's individual ranking, and can hide a strong minority preference entirely.
• Lists are capped at 40. A shortlist that long is not a shortlist, and the question count stops being something anyone will finish.

Common mistakes

Trying to judge availability and appeal in the same pass

Rank on appeal alone, then check availability for the top few. Mixing the two produces an order that reflects neither, and the tool cannot tell you it has happened.

Treating the winner as the decision

It is the name you preferred most often against the others on your list. That is a strong input to a decision and not the decision itself — the top two are usually close enough that other factors should settle it.

Ranking thirty names

Cut to the ones you would actually be happy with first. Ranking names you have already rejected costs questions and changes nothing.

Practical tips

• Run it once yourself before running it with the team. Knowing your own order makes it much easier to tell whether the group's order is a compromise or a genuine consensus.
• Check the flagged inconsistencies before committing. A loop involving your top pick means the shortlist is not settled, whatever the ranking says.
• Run the top three through a confusion check before deciding — a favourite that is easily mixed up with a competitor is not a favourite.

Troubleshooting

The winner is not what I expected

That is usually the point — the method removes the anchoring that makes whichever name you thought of first feel best. If it still looks wrong, run the consistency check: an unexpected winner sometimes means your answers contained a loop.

I want to add a name partway through

Start over with the fuller list. Inserting into a partly finished sort would need the skipped comparisons that were never asked, so there is no honest way to add one mid-run.

Privacy and your data

Your names never leave the browser. They are not uploaded, not stored and never written into the page address — which matters here more than on most tools, because an unlaunched brand shortlist is commercially sensitive and a URL is easy to paste somewhere it should not go. Analytics records only that the tool ran and how many names were in the list, never the names or the resulting order. Start over clears everything immediately.

Frequently asked questions

Why not just score each name out of ten?

Because scores drift. The ninth name gets judged against whatever you have just seen rather than against the first, and people rarely use the scale consistently across a session. A comparison is a single judgement between two things in front of you at the same moment, which is a much more reliable thing to ask for.

How many questions will it ask?

About n log n. Six names take roughly 10 questions, ten take about 23, and twenty about 64 — against 15, 45 and 190 for comparing every pair. The saving comes from never asking a question whose answer your earlier answers already settled.

Does this check whether the name is available?

No. It ranks the names you give it against each other and knows nothing else about them. Trademark registers, company registers and domain availability are separate searches, and nothing here should be read as evidence about any of them.

Can a group use it together?

Yes, and it works well for that — one comparison at a time keeps the discussion specific instead of abstract. Bear in mind that what comes out is the group's answers on the day. It will not show you that two people disagreed strongly about the winner, only where the room landed.