Most teams don’t struggle to make video ads anymore. They struggle to make ads people actually like enough to watch, trust, and act on.
That matters because “liking” isn’t a fluffy metric. In performance environments (Shopify, Amazon, TikTok Shop, Meta), liking is often the difference between:
- a thumb-stopping hook that earns you cheap reach
- and a skip that forces you to buy attention with higher bids
A research team across King’s College London, Erasmus University Rotterdam, University of Amsterdam, and University of Pennsylvania used fMRI brain scans to study what predicts video ad liking. They didn’t just ask people what they preferred. They measured real-time brain activity while 113 people watched 85 video ads, then mapped which creative elements correlated with subconscious liking.
The punchline is refreshingly practical:
People like video ads that start emotional, end logical, evoke empathy, use dialogue (voice + text), and engage both sight and sound.
If you’re producing short-form ads, UGC, product demos, or shoppable videos with AI tools, this is a useful blueprint because it tells you what to scale toward.
Let’s translate the findings into an operator-ready playbook for Shopify, Amazon, TikTok Shop, and Meta commerce.
The 10-second rule: your first job is emotional clarity
The study found the first 10 seconds were the most important in predicting whether people would like the ad.
Not “the first 3 seconds” (the common advice). Ten seconds. Long enough to set a scene, establish a person, and create a feeling.
What “emotional start” actually means in commerce video
This doesn’t mean you need a cinematic backstory. It means the opening should quickly signal:
- a relatable problem
- a social situation people recognize
- a feeling (frustration, embarrassment, stress, guilt, relief, pride)
Example hooks that fit the research:
- “I used to dread taking product photos because…”
- “If your skin does this by 2pm, you’re not alone.”
- “When your package arrives and it’s… not what you expected.”
- “POV: you’re trying to meal prep but your containers always leak.”
For AI video generation, this is good news: emotional clarity is mostly script + performance cues, not expensive production.
The winning structure: emotional beginning, logical ending
The ads people liked most followed a pattern:
- Emotion early (hook with a human moment)
- Logic late (land the solution clearly)
This is the opposite of what many product ads do (feature dump first, vibe later).
How to apply this to shoppable video
Think of it as: feel first, prove second.
- Start: show the pain in a human way
- Middle: show the product in action (not a spec sheet)
- End: make the decision easy (what it is, who it’s for, how to buy)
This maps cleanly to short-form formats:
- TikTok / Reels: hook (emotion) → demo (bridge) → offer/CTA (logic)
- Amazon video: problem context → use case montage → key claims + close
- Shopify PDP video: emotional use case → proof → “what you get” clarity
Empathy is the multiplier (and it’s not the same as “sad”)
A big driver of liking was whether the ad made viewers feel empathy and think about people’s wants and needs (their own or others’).
Empathy can be:
- “that’s me”
- “I know someone like that”
- “I’ve been in that situation”
It’s less about the emotion being positive or negative, and more about being recognizable.
Practical ways to build empathy in UGC-style ads
- Use first-person language: “I,” “my,” “I didn’t realize…”
- Show the “before” moment longer than you think (1–2 extra beats)
- Include a social detail: partner comment, coworker moment, friend reaction
- Use a specific scenario instead of a generic claim
Instead of: “This blender is powerful.”
Try: “I stopped buying smoothies because cleaning my old blender was a nightmare.”
Dialogue beats monologue: use voice + on-screen text
The study found people liked ads more when they included dialogues and narration with both voice and text (captions).
This is especially relevant now because many teams are leaning into silent autoplay and text-only edits. Captions help, but the research suggests the combination matters: hearing + reading reinforces the narrative.
What this means for creators and AI video workflows
If you’re generating ads with AI, don’t stop at visuals. Build a repeatable “dialogue layer”:
- Voiceover (or creator talking to camera)
- On-screen captions that match the spoken line
- Occasional callouts (price, guarantee, bundle) near the end where logic belongs
If you’re using influencer alternatives (in-house creators, paid UGC, AI avatars, founder-led content), the same rule applies: make it feel like a conversation, not a brochure.
Engage sight and sound (yes, sound still matters)
The ads people liked most engaged both:
- Sight: dynamic images, zooms, scene changes
- Hearing: music, sound effects, spoken words
This is a quiet warning to brands producing “pretty” product montages with no audio strategy. Visual polish alone isn’t the point. The brain responds to multi-sensory stimulation because it feels more “real” and immersive.
Quick production upgrades that scale
- Add intentional SFX: click, pour, snap, rip, spray, zip
- Use music to signal emotional tone early, then reduce volume for clarity during the “logic” close
- Cut every 1–2 seconds in the hook section (even if the product is simple)
- Use zooms and reframes to create motion from static footage (especially for Amazon-style product shots)
AI tools can help here, but the operator move is to standardize it: build templates where sound design is not an afterthought.
One counterintuitive finding: avoid “math thinking”
The researchers found:
- Ads that made people evaluate the product (Is it for me?) increased liking
- Ads that made people think about numerical information decreased liking
This doesn’t mean “never mention numbers.” It means: don’t force cognitive load.
Where this shows up in ecommerce ads
- complicated discount math (“buy 2 get 15% then extra 10% at checkout”)
- too many specs too fast
- capacity comparisons without context
- pricing tables inside a 15-second video
Better approach: translate numbers into outcomes.
- “One charge lasts all weekend” instead of “10,000 mAh”
- “Fits a full outfit + shoes” instead of “32L”
- “Stops leaks in your tote” instead of “IPX7”
On Amazon, you still need specs. Just move them to:
- the last third of the video
- simple overlays
- one claim per beat
Platform-by-platform: how to use this without fighting the format
Shopify (PDP video + paid social)
Shopify brands often have the most control over storytelling, but they also tend to over-index on brand polish.
Use the research-backed structure:
- Opening: customer pain in a real setting (UGC style wins here)
- Middle: demo + proof (show it working)
- End: rational close (what it is, who it’s for, guarantee/shipping)
If you’re thinking about how AI changes the creative advantage, this connects directly to Tellos’ view in How AI helps creative Shopify brands stand out: scale doesn’t help if you scale the wrong pattern. You want scale with a creative spine.
Amazon sellers (listing video + Sponsored Brands Video)
Amazon shoppers are already in “logic mode,” but the ad still needs an emotional entry to earn attention.
- Start with the moment of frustration (the thing that made them search)
- Then show the product solving it clearly
- End with simple, non-mathy proof points (warranty, compatibility, what’s included)
Avoid dense spec narration. Let the listing carry the details; let