Methods of quantitative evaluation of a brand name prior to launch
Why 'liking' is a deceptive criterion
The standard questionnaire with the question 'how much do you like this name' works against innovative names. Apple, Yahoo, Urban Decay — all would fail the classic quantitative test for likability and category fit. The reason lies in the perception mechanism: a brief emotional assessment triggers a quick reaction of the 'system 1' type — intuitive, based on familiar patterns. Unfamiliar, abstract names require analytical processing — 'system 2', which works slower and gives low scores on the first wave.
Names that violate the familiar linguistic logic for the category are often perceived worse at first glance, but are better remembered without prompts in the long run. Long vowels in names can subconsciously associate with scale and weight — and you get a tool that the standard likes/fit scale simply fails to detect.
How many respondents are needed for a valid sample
The minimum threshold for quantitative brand testing is 300 respondents, with a comfortable working standard of 500–600. With a sample of 500 people, the margin of error remains within ±4%, and when increased to 1000 respondents, it decreases to ±3%. For legally significant research, for example, when preparing for trademark registration in Russia, a more stringent professional standard applies.
| Context | Minimum sample size | Margin of error |
|---|---|---|
| Marketing testing (working standard) | 300–600 respondents | ±3–4% at 500 |
| High accuracy (strategic decisions) | 1000+ respondents | ±1–2% |
| Trademark legal expertise (Russia) | 500 in 2 cities + 125 in the rest | is determined by the expert methodology |
Saving on the sample size for the sake of speed — a typical mistake: the result on 80–100 respondents statistically proves nothing except the opinion of a narrow group that accidentally fell into the survey.
How to build a questionnaire: the correct sequence of questions
The order of questions solves not less than their formulation. If you show the respondent the name with hints before asking about free recall, you will lose the opportunity to measure the natural salience of the name forever. The correct logic of the questionnaire looks like this:
- Unaided recall — a question without hints: «What names in this category can you recall right now?». Captures what really lives in memory without stimulus.
- Aided recall — the respondent is shown a list of several names, including the one being tested, and asked which of them they recognize. This is recognition, not spontaneous recall — and they must not be confused.
- Description exposure — the name is shown together with a brief product description or positioning and the respondent is asked to evaluate perception: uniqueness, relevance, impression, likelihood of purchase.
- Repeat / delayed measurement — repeated contact with the respondent after 48 hours without warning, to check what really stuck in memory.
Before the main block, screening is mandatory: for products of prolonged use — a question about purchase frequency in the category, for everyday demand products — verification of a recent purchase. Without this filter, people who are not the target audience of the brand at all end up in the sample, and the results are distorted.
What shows the first reaction, and what — memory after 48 hours
The immediate reaction to a name and what will remain in the mind after two days are literally two different studies. The Ehrenberg-Bass Institute insists that the true measure of a name's effectiveness is delayed recall after 48 hours, not the rating given immediately after exposure. The mechanism is simple — short-term memory records everything that was recently presented, while long-term memory filters out the noise and retains only what resonated with existing associations or stood out from the competition.
This practically means a second contact with the same group of respondents after two days, without warning about a repeated survey, with the same unaided recall question. The difference between the 'immediately' and 'after 48 hours' indicators is the real forecast of how the name will live on the market, not in the laboratory conditions of a focus group.
Monadic, sequential and paired comparison: what to choose
The method of presenting names affects the result no less than the questionnaire itself.
| Method | How it works | When to apply |
|---|---|---|
| Monadic testing | Each group of respondents is shown only one name, without comparison with others | When it is important to exclude the effect of comparison and evaluate the name 'as it is' |
| Sequential | The respondent sees several names in turn, evaluating each separately | When the budget is limited and you need to test several options on one sample |
| Pairwise comparison | The name is evaluated relative to others, pairwise | When you need to exclude distortion due to re-evaluation or under-evaluation during binary choice |
In addition to basic scales, it is advisable to include an assessment of cognitive load — how easy it is to pronounce and spell the name letter by letter. This is not a secondary parameter: studies show that lower cognitive load correlates with a higher level of trust in the brand.
How to read the results when the data contradict each other
The most common managerial mistake is to discard a name merely because it didn't score high on 'likes.' If a name is poorly rated on likability but confidently wins on differentiation and delayed recall after 48 hours, this is not a contradiction but a normal pattern for a truly new name. The focus should not be on a single metric but on a combination of three indicators: visibility (spontaneous awareness), differentiation (how much the name differs from competitors), and recall without cues after two days.
If all three indicators are low — the name is weak, and there is nothing to argue about. If only the initial sympathy is low, while differentiation and delayed recall are high — you are looking at a candidate that will work after market entry, when the audience has time to get accustomed. The reverse situation occurs even more often and is much more dangerous: the name is liked at first glance, easily gets high scores on the fit scale, but fails in delayed recall. This is a classic sign of a 'safe' but forgettable name — it does not create friction, and therefore does not leave a trace in memory.
The practical rule is simple: the decision to launch a name is not based on the average score, but on the profile of several metrics taken at different times from the same group. One measurement immediately after the show is a snapshot of mood, not a forecast of market behavior.
What to do with contradictory results in practice
If the test yielded a mixed picture, there is a working algorithm for making a decision:
- Check whether the low 'liking' is related to an unfamiliar linguistic approach for the category — this is a sign of potential, not a failure.
- Compare the delayed recall of the tested name with that of category leaders — if the difference is within the margin of statistical error, the name is viable.
- Assess the cognitive load separately: if the name is hard to pronounce or write down, this is a standalone risk, independent of other metrics.
- Do not make a decision based on a sample of fewer than 300 people and do not rely solely on the first wave of polling without a follow-up measurement after 48 hours.
Quantitative naming testing is not a formality before signing the brief for design, but the only way to distinguish a name that will truly stick in the market's memory from a name that everyone likes in negotiations but dissolves in informational noise already within a week after launch.
Why 'liking' is a deceptive criterion
The standard questionnaire with the question 'how much do you like this name' works against innovative names. Apple, Yahoo, Urban Decay — all would fail the classic quantitative test for likability and category fit. The reason lies in the perception mechanism: a brief emotional assessment triggers a quick reaction of the 'system 1' type — intuitive, based on familiar patterns. Unfamiliar, abstract names require analytical processing — 'system 2', which works slower and gives low scores on the first wave.
Names that violate the familiar linguistic logic for the category are often perceived worse at first glance, but are better remembered without prompts in the long run. Long vowels in names can subconsciously associate with scale and weight — and you get a tool that the standard likes/fit scale simply fails to detect.
How many respondents are needed for a valid sample
The minimum threshold for quantitative brand testing is 300 respondents, with a comfortable working standard of 500–600. With a sample of 500 people, the margin of error remains within ±4%, and when increased to 1000 respondents, it decreases to ±3%. For legally significant research, for example, when preparing for trademark registration in Russia, a more stringent professional standard applies.
| Context | Minimum sample size | Margin of error |
|---|---|---|
| Marketing testing (working standard) | 300–600 respondents | ±3–4% at 500 |
| High accuracy (strategic decisions) | 1000+ respondents | ±1–2% |
| Trademark legal expertise (Russia) | 500 in 2 cities + 125 in the rest | is determined by the expert methodology |
Saving on the sample size for the sake of speed — a typical mistake: the result on 80–100 respondents statistically proves nothing except the opinion of a narrow group that accidentally fell into the survey.
How to build a questionnaire: the correct sequence of questions
The order of questions solves not less than their formulation. If you show the respondent the name with hints before asking about free recall, you will lose the opportunity to measure the natural salience of the name forever. The correct logic of the questionnaire looks like this:
- Unaided recall — a question without hints: «What names in this category can you recall right now?». Captures what really lives in memory without stimulus.
- Aided recall — the respondent is shown a list of several names, including the one being tested, and asked which of them they recognize. This is recognition, not spontaneous recall — and they must not be confused.
- Description exposure — the name is shown together with a brief product description or positioning and the respondent is asked to evaluate perception: uniqueness, relevance, impression, likelihood of purchase.
- Repeat / delayed measurement — repeated contact with the respondent after 48 hours without warning, to check what really stuck in memory.
Before the main block, screening is mandatory: for products of prolonged use — a question about purchase frequency in the category, for everyday demand products — verification of a recent purchase. Without this filter, people who are not the target audience of the brand at all end up in the sample, and the results are distorted.
What shows the first reaction, and what — memory after 48 hours
The immediate reaction to a name and what will remain in the mind after two days are literally two different studies. The Ehrenberg-Bass Institute insists that the true measure of a name's effectiveness is delayed recall after 48 hours, not the rating given immediately after exposure. The mechanism is simple — short-term memory records everything that was recently presented, while long-term memory filters out the noise and retains only what resonated with existing associations or stood out from the competition.
This practically means a second contact with the same group of respondents after two days, without warning about a repeated survey, with the same unaided recall question. The difference between the 'immediately' and 'after 48 hours' indicators is the real forecast of how the name will live on the market, not in the laboratory conditions of a focus group.
Monadic, sequential and paired comparison: what to choose
The method of presenting names affects the result no less than the questionnaire itself.
| Method | How it works | When to apply |
|---|---|---|
| Monadic testing | Each group of respondents is shown only one name, without comparison with others | When it is important to exclude the effect of comparison and evaluate the name 'as it is' |
| Sequential | The respondent sees several names in turn, evaluating each separately | When the budget is limited and you need to test several options on one sample |
| Pairwise comparison | The name is evaluated relative to others, pairwise | When you need to exclude distortion due to re-evaluation or under-evaluation during binary choice |
In addition to basic scales, it is advisable to include an assessment of cognitive load — how easy it is to pronounce and spell the name letter by letter. This is not a secondary parameter: studies show that lower cognitive load correlates with a higher level of trust in the brand.
How to read the results when the data contradict each other
The most common managerial mistake is to discard a name merely because it didn't score high on 'likes.' If a name is poorly rated on likability but confidently wins on differentiation and delayed recall after 48 hours, this is not a contradiction but a normal pattern for a truly new name. The focus should not be on a single metric but on a combination of three indicators: visibility (spontaneous awareness), differentiation (how much the name differs from competitors), and recall without cues after two days.
If all three indicators are low — the name is weak, and there is nothing to argue about. If only the initial sympathy is low, while differentiation and delayed recall are high — you are looking at a candidate that will work after market entry, when the audience has time to get accustomed. The reverse situation occurs even more often and is much more dangerous: the name is liked at first glance, easily gets high scores on the fit scale, but fails in delayed recall. This is a classic sign of a 'safe' but forgettable name — it does not create friction, and therefore does not leave a trace in memory.
The practical rule is simple: the decision to launch a name is not based on the average score, but on the profile of several metrics taken at different times from the same group. One measurement immediately after the show is a snapshot of mood, not a forecast of market behavior.
What to do with contradictory results in practice
If the test yielded a mixed picture, there is a working algorithm for making a decision:
- Check whether the low 'liking' is related to an unfamiliar linguistic approach for the category — this is a sign of potential, not a failure.
- Compare the delayed recall of the tested name with that of category leaders — if the difference is within the margin of statistical error, the name is viable.
- Assess the cognitive load separately: if the name is hard to pronounce or write down, this is a standalone risk, independent of other metrics.
- Do not make a decision based on a sample of fewer than 300 people and do not rely solely on the first wave of polling without a follow-up measurement after 48 hours.
Quantitative naming testing is not a formality before signing the brief for design, but the only way to distinguish a name that will truly stick in the market's memory from a name that everyone likes in negotiations but dissolves in informational noise already within a week after launch.
Recent articles
MegaFon, BMW, VTB and more 240+ companies. See how it looks in the showcase reel:
Our branding agency is approached to order:
Branding agency services
