A/B testing of new brand elements before launch
Most brand testing guides suggest showing the logo to 20 respondents and asking "do you like it?" This is not a test, but an unstructured focus group: there is no sample calculated by power formula, no point at which the test can be considered completed, no division of stages by error cost. Below is a process with specific thresholds, not intuition of a designer or marketer.
What to test first — and why order matters
Brand elements differ by one criterion: how expensive it is to correct an error after launch. Hence the order of testing — from the most irreversible to the most flexible.
- Brand name and nomenclature — an error is almost impossible to fix without losing recognizability and legal costs for rebranding, so it is tested first and on the widest possible sample.
- Logo and visual identity — affects first impression and trust, but is updated relatively painlessly even years after launch.
- Slogan and key message — tested together with the logo, usually via 5-second impression tests.
- Packaging and touchpoints — requires real traffic and physical environment, not abstract evaluation on a screen.
- Tone of communication (playful vs sophisticated and other polarities) — tested only on live advertising campaigns, because declared audience preferences and reaction to real advertising diverge systematically.
Tone is the most treacherous variable precisely because cabinet research fails to capture it. Supermetrics launched a full-fledged A/B test consisting of nine experiments (YouTube and Facebook video ads, static Facebook and Google ads, brand lift and ad recall tests), comparing the 'playful' and 'sophisticated' brand variants,
Most brand testing guides suggest showing the logo to 20 respondents and asking "do you like it?" This is not a test, but an unstructured focus group: there is no sample calculated by power formula, no point at which the test can be considered completed, no division of stages by error cost. Below is a process with specific thresholds, not intuition of a designer or marketer.
What to test first — and why order matters
Brand elements differ by one criterion: how expensive it is to correct an error after launch. Hence the order of testing — from the most irreversible to the most flexible.
- Brand name and nomenclature — an error is almost impossible to fix without losing recognizability and legal costs for rebranding, so it is tested first and on the widest possible sample.
- Logo and visual identity — affects first impression and trust, but is updated relatively painlessly even years after launch.
- Slogan and key message — tested together with the logo, usually via 5-second impression tests.
- Packaging and touchpoints — requires real traffic and physical environment, not abstract evaluation on a screen.
- Tone of communication (playful vs sophisticated and other polarities) — tested only on live advertising campaigns, because declared audience preferences and reaction to real advertising diverge systematically.
Tone is the most treacherous variable precisely because cabinet research fails to capture it. Supermetrics launched a full-fledged A/B test consisting of nine experiments (YouTube and Facebook video ads, static Facebook and Google ads, brand lift and ad recall tests), comparing the 'playful' and 'sophisticated' brand variants,
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
