What Makes a Video Ad "Good": Defining Quality Beyond Gut Feeling

Ask five people whether an ad is good, and you'll get five answers — most of them based on whether they personally liked it. This is why creative review by committee produces mediocrity. Personal preference, stated confidently, gets mistaken for objective evaluation.

Liking vs. effectiveness

People conflate liking an ad with the ad being effective. They're different things, weakly correlated. An ad can be intensely likeable and entirely ineffective — the viewer enjoys it, remembers the entertainment, and has no idea what brand was being advertised. An ad can be grating and highly effective — the viewer doesn't enjoy it, but the message gets through and influences behavior.

The neuroscience supports this distinction. The Chan et al. (2024) fMRI study of 113 participants watching 85 video ads found that emotional response predicted ad liking within the first three seconds, but liking predicted essentially nothing about behavior. Social cognition — whether the viewer connected the ad to their own situation — was the stronger predictor of downstream action.

When evaluating creative, separate the question "do I like this?" from "will this work?" They're different questions with different answers.

Four dimensions of ad quality

Clarity. Can a viewer who has never heard of the product describe what it does after watching once? If not, the ad has a clarity problem. This is the most common failure in creative — the ad is clever, beautiful, or emotional, but the message is missing.

Clarity is testable: show the ad to someone outside your industry. Ask "what does this product do?" If their answer doesn't match your intent, the ad isn't clear enough, no matter how good it looks.

Distinctiveness. Would the viewer recognize this ad as yours, even without the logo? If the brand could be swapped for a competitor without anyone noticing, the ad isn't distinctive. It's generic creative wearing your logo.

Distinctiveness can come from visual identity, tone, format, casting, or a hundred other creative choices. What matters is that something about the ad is unmistakably yours — something a competitor couldn't or wouldn't do.

Emotional engagement. Does the ad make the viewer feel something — anything — at an intensity above background? The specific emotion matters less than the intensity. Mildly positive is not enough. Mildly amusing is not enough. The ad needs to move the viewer's emotional state from wherever it was to somewhere noticeably different.

This is where AI-based testing tools are most useful. They can quantify emotional engagement in ways that human intuition can't, because humans confuse their own emotional response with the target audience's likely response.

Message retention. After one viewing, what does the viewer remember? If they remember the joke but not the brand, the ad is entertaining but not effective. If they remember the brand but not what it does, the ad is branding without messaging. If they remember both, the ad is doing its job.

The stakeholder problem

The biggest obstacle to objective creative evaluation is stakeholder preference. A VP likes the color blue. A brand manager hates voiceover. A CEO wants the logo bigger. These preferences, stated as opinions, become requirements. The ad gets optimized for internal approval rather than external performance.

The fix: test creative with the target audience before showing it to stakeholders. When the data says a creative approach works, stakeholder preference becomes less relevant. The conversation shifts from "I like this" to "the audience responds to this." That's a better conversation.

A pre-review checklist

Before reviewing any ad creative, answer these questions without looking at the ad again:

What does the product do? (clarity)
What feeling did the ad leave you with? (emotional engagement)
What's the one thing you'll remember tomorrow? (message retention)
Could a competitor have made this ad? (distinctiveness)

If you can't answer all four clearly, the ad has a problem. Which problem determines what you fix.