You have three favourites, and everyone on the team prefers a different one. But voting for the nicest name will not tell you whether customers can spell it after hearing it over the phone, or whether it suggests an entirely different service. Before making your final choice, try the names in a few specific situations.
The aim of a small test is to uncover problems and understand reactions from people in your target audience. It will not guarantee brand success or legal availability. However, it gives you better evidence than asking friends, “Do you like it?”
First, decide what you need to find out
Test a shortlist, such as two or three names. This is a practical recommendation for a manageable conversation, not a universal research rule. If you have twenty options, first filter them against your brief. See the guide to developing and choosing a company name.
For each candidate, write down a specific uncertainty:
- Can customers read the invented word?
- Will they spell an English name correctly when they only hear it?
- Does the name suggest a different sector?
- Does it convey the tone you need, such as straightforward, playful or calm?
- Can people recall it after a short while without a prompt?
Also decide in advance what would count as a serious problem for you. For a service sold mainly over the phone, correct spelling may matter more than for a product whose name people usually read on packaging. Your criteria should reflect actual use.
Who to recruit and how many people you need
Look for people who resemble your future customers. If you offer a service to small business owners, reactions from classmates with no business experience may answer a different question. Relatives or colleagues can help you check whether the instructions are understandable, but should not automatically stand in for your market.
Selecting participants according to the research question is also recommended in GOV.UK's recruitment guidance. Here, we apply its general principles to a small naming exercise, rather than treating them as a complete methodology for a representative brand survey.
For an initial indication, you can start with a few individual interviews. Record whom you recruited and how they fit the target audience. If you sell to substantially different groups, do not treat all responses as one undifferentiated set.
No universal number of participants will confirm that a name works. The scope needed depends on the question and how participants are selected. For statistical comparisons between options ahead of a larger investment, you need a separate research design.
Prepare the same conditions for every name
Initially show the names as plain text in the same font and size. Do not give one option a beautiful logo and another just a line in a message. You would be measuring reactions to different presentations.
Prepare the same short service description and the same questions. Vary the order of candidates between participants. Pew Research Center describes the effects of question wording and order. Alternating the order is a practical precaution here; it does not turn a small test into a statistically balanced experiment.
Also decide what someone should encounter first. If you want to test spelling from hearing, they must not read the name on a screen beforehand. If you start with reading, later spelling no longer measures their first encounter with the name. You can use different participants for different first impressions and record that distinction.
1. Say the name and ask the participant to write it down
Say the candidate at a natural pace, without spelling it out. Ask: “Please write down the name you just heard.”
Record the result and any follow-up questions. Did they ask you to repeat the name? Were they unsure whether it was one word? Did they add a letter or write a Czech spelling of an English word?
Do not correct them during the task. Afterwards, you can show the intended spelling and ask what was unclear. You are looking for an obstacle customers might encounter when recommending your service to someone else.
If you have a proposed domain, run a separate task with the address too. Pronouncing the brand and dictating the domain may not be equally easy. Ask participants only to write the address down; do not send them to an unchecked third-party website.
2. Show the name and ask them to read it
Ask: “How would you say this name aloud?” Watch for different pronunciations and hesitation. For someone who has already heard the name, note that this is a repeat encounter.
A different pronunciation need not be a reason to reject the name. What matters is whether it complicated communication and whether you can accommodate it in practice. However, for a name based on a particular play on words, it may show that the intended meaning remains hidden.
3. Find out the first associations without explaining
Start with open questions:
- “What first comes to mind when you see this name?”
- “What would you expect it to offer?”
- “Does it remind you of anything you already know?”
Do not ask “Does it sound trustworthy?” or “Isn't it too childish?” if you want to capture a spontaneous impression. Such questions suggest both a topic and a possible answer. GOV.UK's interview guidance recommends avoiding leading questions.
Then show your prepared business description and ask: “How well does the name fit this offer? What is unclear about the connection?” The name does not have to explain the entire product on its own. But you need to know whether the accompanying description helps or whether the mismatch remains.
4. Return to the name without a prompt
After the same short, unrelated activity, ask participants to recall the names they were shown. Do not leave the list visible. Also record the approximate interval and how many times they encountered the names.
Distinguish unaided recall from recognising a name in a list. Someone may recognise the right option among several choices but be unable to say it independently. Both tasks have uses; do not treat their results as the same thing.
A short exercise during an interview does not show what a customer will remember a month later. Use it as information about the particular situation you set up.
How to record and assess responses
Create a record for each participant:
- how they fit the target audience;
- the order of candidates and how they were first introduced;
- their actual spelling and pronunciation;
- first associations, ideally in their own words;
- reactions after the offer was explained;
- recall after an interval;
- your conclusion, kept separate from the original response.
Look for recurring obstacles and their consequences. If someone says “I would expect accounting”, do not immediately rewrite that as “the name is bad”. Compare the response with the actual offer and the other records.
For a small sample, report specific counts alongside the total number of participants. Do not turn them into an estimate of how the entire market behaves. Nor should a simple total score hide a serious obstacle in a critical use case.
Finally, decide whether to keep the name, adjust it and test again, or reject it. If you change the spelling, for example, include new people in the next exercise. The original participants already know the explanation.
A positive reaction does not replace checking the name
User testing and checking rights address different questions. People may like a name that resembles an existing brand. Before your final decision, complete the company, domain and trademark checks.
If the test shows that you need different candidates, generate more ideas on OverNazev.cz and return to the specific problems in your records. A precise note such as “people spelled the name three different ways after hearing it over the phone” helps a new brief more than a general request for “something better”.