Creative testing mistakes to avoid

We’ve covered what you SHOULD do. Here’s what you SHOULDN’T do in your creative testing research. 

Don’t test too many assets

If you’re asking people to compare several creatives in a sequential survey, fatigue can quickly set in. Don’t try to test more than five or six ideas in a single survey—ideally, you should aim for just two or three. 

Remember: the type of creative you’re testing will affect your respondents’ reactions. They could find multiple video ads overwhelming, while showing them two static ads would be much less stressful. And regardless of asset type, keep an eye on how busy the visuals are—even a static ad can be overwhelming if it has lots of visual elements or colors, so keep in mind what is impactful for different types of creatives. 

Don’t ask too many questions

You should also limit the number of questions you ask respondents in creative testing so they don’t lose focus. 

For example, if you were testing three different creative assets, ideally don’t ask them more than five questions per asset. If you do, their attention may wander. 

Pro tip: A good way to judge how many questions you need is to look at the whole survey. You can get away with asking a few more questions per asset if, overall, your survey isn’t too long. With Attest you have a limit of 40 questions per survey, so you have the freedom to go as deep with your research as you need to.

Only focus on the questions that really matter. If you have a predefined idea of what your success metrics look like before you conduct the iterative creative testing, you’re more likely to get the data you need to inform your decision-making. 

Don’t test creatives that are too similar to each other

If you test three very similar creatives in a sequential test, you’re unlikely to find a clear winner. Also, don’t test multiple similar ideas alongside one that is wildly different—they’ll see that the different one is unusual, and this might bias their responses. 

If you want to test very similar ideas, use monadic testing (one asset per survey) so respondents just focus on one asset and aren’t aware of the other ideas that are being tested.