Getting consumer feedback early and often during creative development leads to more effective campaigns. An iterative research loop might include:

• Starting with a consumer insight that’s relevant to your category
• Developing a campaign idea that brings the brand’s take on that insight to life, and can work across time, media and markets
• Testing storylines for video and audio content to understand what resonates, what doesn’t and what to build on in finished ads, and iterating with rough versions of static or digital ads
• Refining unfinished and finished ads so they resonate and cue the brand across every channel you invest in

Each round of research is most useful when a change is likely to affect how people experience the ad. This article covers how to decide when to retest, how to choose category drivers and how to compare executions.

Why small changes rarely show up in results

In real life, people don’t pay full attention to ads. Media space can be bought, but attention is earned, and even ads that earn it aren’t remembered in every detail. People remember elements: certain characters, a rough story, maybe the music, ideally the brand, and a feeling or message they took away.

So unless a change makes the ad more likely to grab attention, tell a clearer or more relevant story, or give the brand a stronger role, research results are unlikely to show a meaningful difference, either in how the ad is received or in why people like or dislike it.

Example: changing the end card

End cards usually aren’t connected to the story. They appear as a closing frame, after the ad is “over” for most viewers, and take up little of what people remember. Unless the end card ties into the story, for example by reusing a moment or character from the ad, few people notice it. In most cases, switching out an end card doesn’t change research results.

Deciding whether a change is worth retesting

Focus on how a change might affect the way people experience and remember the ad, not on the size of the change. Pay particular attention to the role of the brand, comprehension, emotional connection and distinctiveness. If you answer yes to any of the questions below, it’s likely worth researching the revised edit or edits.

Story meaning and structure

• Are your edits likely to change the story people recall from the ad, or improve their understanding of its story or meaning? Did your previous research pinpoint an issue in how people remember or understand the story that the revised edit solves?
• Especially for problem and resolution ads, have you changed the story to take people on a better emotional journey and create a strong enough peak end?
• Have you changed the ad length significantly? This could mean extending it to tell a clearer story or add moments of impact, or shortening it so some elements of the story are no longer there.

Scene selection

• Have you removed scenes that took up space but weren’t resonating or being noticed? For example, scenes with no lift in moment-by-moment emotional response, no mention in story recall, or a mention as a dislike.
• Starting without previous results? Research a few edits that are clearly different from each other, rather than small adjustments to a “filler” scene that’s unlikely to affect the meaning or emotional reaction.
• Have you added key scenes to improve understanding, or to enhance the story or entertainment value?
• Have you adapted key scenes in a meaningful way? Changes to how people interact or the emotions they display may have an impact, as can changes to the role of the brand or how people see it within the ad. A slightly different camera angle or a small font size change within a longer video is less likely to affect overall response.

To test small adjustments within a scene, such as font size, research that frame on its own.

Audio

• Have you significantly changed the voiceover in a way that could affect the story or how it’s received? For example, new voices, a cut voiceover or added clarity.
• Have you changed the music to a track that shifts the emotion or mood, or reinforces the story arc?

Role of the brand

• Do any of the edits change the role of the brand and make it more likely to be noticed during the ad?
• Have you integrated more distinctive assets?
• Have you added something to clarify how the story relates to the brand?
• Have you given the brand a role in the more memorable moments of the ad?

Characters

• Have you changed any of the main characters in the story, such as their personality or appearance?

Messaging and associations

• For a functional ad, have you adapted how key messages or reasons to believe are delivered, in response to low message takeout or credibility?
• Have you changed how an association or feeling is conveyed, or what the ad aims to convey? Check your results against your category drivers.

Use your previous results as a guide

Your earlier research results can show how likely a change is to make a difference:

• Did you remove scenes that weren’t creating emotional reactions in the moment-by-moment responses?
• Did you add scenes intended to evoke emotion where it was lacking, such as something to make people laugh or feel love? Did you change the story arc because a problem wasn’t resolved positively enough?
• Did story playback or dislikes show that people didn’t understand the story, and have you adapted it to fix that?
• Did you identify a branding issue and restructure the story, distinctive assets or brand presence into the more memorable parts of the ad?

If your changes are small

We recommend skipping further research. For a change like a new end frame, test the options on their own with Screen It to see which is most compelling.

Choosing category drivers

Category drivers are the reasons people choose within a category: the needs, benefits and occasions that decide which brand gets bought. Amplify measures how strongly an ad links your brand to these drivers. This shows whether your creative builds the associations that matter for growth in the category, not just whether people like the ad.

What makes a good driver statement

• It’s about the category, not the brand: It should be true of any brand in the category, such as “is refreshing,” not “has the iconic blue can.” Brand-unique assets belong in a separate distinctiveness measure.
• It’s a real reason to buy: Choose statements because they matter to choice, based on U&A data, category entry point (CEP) work or driver analysis, not because they match the brand’s messaging.
• It holds one idea: Avoid double-barreled statements like “tasty and healthy,” because you can’t tell which half drove the response.
• It uses consumer language: Keep it short, concrete and easy to translate. “Good to share with friends” works better than “facilitates social connection.”
• It’s neutral, not evaluative: Leave out “the best,” “better than” and other comparatives. These measure overall appeal, not a driver.
• It describes the brand, not the ad: Words like “funny” or “cool” describe the ad, not what it conveys about the brand.

What a balanced set covers

Aim for roughly 8 to 12 statements across:

• Functional benefits: taste, refreshment, ingredients, satisfies hunger
• Emotional and social benefits: fun, a treat for me, bringing people together
• Occasions or CEPs: when, where and with whom it’s used, such as “for watching sports at home” or “an afternoon pick-me-up”
• Practical factors (optional): value, easy to find, comes in sizes that suit me

Comparing executions

When you’re choosing between executions:

1. Use the stimuli-to-stimuli comparison view, not the ad-to-norm comparison view.
2. Check for significant differences in key measures: Focus on the measures that feed the sales and brand impact scores. If one ad is significantly better on several key measures, progress it and look at how to optimize it further. If there are no significant differences, the ads perform similarly, and choosing between them takes further diagnosis using open responses and moment-by-moment emotional reaction.
3. Diagnose the reasons behind any differences:
   • Overall emotion: Which edit conveys the desired emotion more?
   • Emotional moments: Which edit best follows the intended emotional journey, with more or stronger peaks in key emotions? Which leaves people feeling most positive at the end?
   • Story playback: Is one ad more clearly understood than the other?
   • Unaided messaging: Is one ad more clearly conveying the intended associations?
   • Category drivers: Which edit more strongly links the brand to the drivers that matter in your category?
   • Open-ended responses: Do respondents often mention particular moments, characters or themes, positively or as dislikes, that appear in one edit but not the other? Does one version deliver a more memorable story?
4. Don’t compare percentile scores: A change in percentile score doesn’t necessarily reflect a significant difference. Percentiles give an at-a-glance sense of how an ad performed (great, good, OK or poor). On the stimuli comparison view, significance testing between ads uses the absolute score, so use that to confirm whether a difference is meaningful.