Name testing: how to validate a brand or product name before launch

A product can have the right features, the right price, and the right market, and still lose people at the name.

Someone can't spell it, can't say it out loud, or reads a meaning into it that nobody intended, and the decision to buy never gets that far. By the time that shows up in the data, the name is already on packaging, in the app store, and filed with the trademark office.

Name testing is how you catch that before it happens.

This guide covers what the method does and doesn't do, the attributes worth measuring, how to set up the stimulus and sample so the results hold up, the failure modes that quietly wreck a test, and a step-by-step process for running one and reading the results.

Name testing is a market research method that measures how a defined target audience reacts to candidate names for a brand, product, service, or feature before that name reaches the market.

In marketing, the term means survey-based measurement of naming options, not software naming conventions and not the Boston Naming Test used in neuropsychology.

The scope is deliberately narrow. Name testing does not invent names. Name testing picks up at the shortlist and answers one question: which of these names performs best with the people who will actually buy?

Generating candidates is a creative exercise, and if that's where you are, start with the guidance on choosing a product name and come back once you have a shortlist. 

It's also not concept testing. A concept test evaluates the idea, the benefit, and the offer. A name test holds the idea constant and varies only the label attached to it. Both sit alongside product testing, which evaluates the thing itself after people have used it.

So treat name testing as measurement, not naming strategy. A name test tells you what a candidate does to comprehension, appeal, and buying intent. What the name should stand for in the first place is decided upstream, before any survey goes into the field.

A name is one of the last launch variables that's still cheap to change and one of the first that becomes expensive to fix. Once it's on packaging, in the app store, on the trademark register, and in paid search, changing it means redoing all of it.

Here's the outside evidence on what's at stake.

Beyond the figures, three business risks show up again and again when a name goes untested.

A name test that only asks "which do you prefer" produces a popularity contest, not a decision. Measure the attributes that actually predict performance, and keep the same battery across every candidate.

Open text on associations is where the surprises live. A name can win on appeal and still carry a meaning in one market that rules it out entirely.

Respondents can only react to what you show them, which makes stimulus preparation the part of a name test that most often gets rushed. Four decisions matter.

Give every candidate the same short product description, one or two sentences, and keep the word count identical across groups. A name shown with a richer description will score better on fit for reasons that have nothing to do with the name.

Name the category before you show the name. "A new energy drink called X" and a bare "X" produce different comprehension scores, and only the first reflects how the name will be encountered in the wild.

Text-only isolates the name, which is what you want when the design isn't locked. Once you're testing a lockup, you're partly running a logo testing exercise, and the two effects become hard to separate. Keep them in separate studies, or use logo design testing templates for the visual question.

If a name is invented, non-English, or ambiguously spelled, record a two-second clip and play it. Without audio you're measuring how people guess at the name rather than how they'll hear it in an ad.

Practical design comes down to four calls. Show each respondent one name and one name only, so nobody scores a candidate by comparing it to the last one they saw. Give each name its own cell with the same question set. Include a control name, either the incumbent or a real competitor already in market, because a score with nothing to sit next to isn't interpretable. Randomize assignment so cells stay comparable on demographics.

Cell count follows from the shortlist you can afford to field, not from theory. If you're launching in more than one language, each market needs its own cells. For the reasoning behind one-at-a-time exposure and how it compares with other approaches, see the breakdown of monadic survey design.

Four categories of resource do the work, and most teams already have two of them sitting unused.

If naming decisions sit inside a wider market research program, wire the name test into that calendar rather than treating it as a one-off favor to the brand team. Reusing the same attribute set across launches is what turns a single result into a benchmark you can compare against next year.

A name test is a small, cheap study that protects a decision you'll be living with for years. Pick your attributes, prepare the stimulus with care, give every name its own cell and a control to sit beside, and read the results as ranges rather than a leaderboard. The teams that get this right aren't the ones with the best instincts. They're the ones who checked.

Explore the product