In the last 18 months, Top-of-Search (TOS) video ads – especially Sponsored Brands Video – have gone from novelty to frontline weapon. On mobile, they’re often the first thing a shopper sees under the search bar: a full width, auto playing video that interrupts the usual grid of thumbnails.
That placement is incredibly powerful and incredibly unforgiving. Most shoppers will give your video about two seconds before they decide to either keep scrolling or pay attention. If those two seconds are wasted, the rest of your beautifully produced 30 seconds might as well not exist.

This article looks at how TOS video works, what those first moments need to do, and how to use A/B testing to get your creative to a point where it consistently earns clicks instead of leakage.
Why does TOS video matter so much now?
A TOS video ad is Amazon’s answer to the modern discovery habit. Shoppers don’t just read; they watch and skim. Video at the top of search lets you break the pattern of lookalike thumbnails, show your product in use instantly, and deliver a mini sales pitch before the shopper ever reaches a detail page.
Because video takes up more space and movement than static ads, it tends to pull disproportionate attention. When it’s relevant and well-executed, that means higher click-through and more qualified traffic. When it’s off the mark, it becomes an expensive way to tell people “You’re in the wrong place”.
The crucial detail: video ads default to autoplay, muted, and often fill a big chunk of the mobile screen. Shoppers don’t choose to watch your video; they run into it. That means the opening frame – and the next second or two – are all that really determine whether they stay.
The first two seconds: what they must do
You don’t have time for logo animations, drone shots of scenery, or slow reveals. In the first two seconds, your video needs to answer three questions in the shopper’s head: What is this? Is it for me? Is it worth a tap?
If those answers are not obvious visually, the scroll continues.
Strong openings usually share these traits: the product is clearly visible, not tiny in the corner; the use-case or outcome is visible (someone actually using it, or a clear before/after); and the category is instantly recognisable even muted (e.g. “oh, that’s a phone case / water bottle / posture corrector”).
A weak opening often looks like this: logo on a neutral background, generic lifestyle shot where the product is hard to spot, or a busy scene where nothing stands out.
As a rule of thumb: if a shopper watching your video for two seconds on mute can’t tell what you sell and what it does for them, your hook needs work.
Structuring the rest of the video (in 15–30 seconds)
Once you’ve nailed the first two seconds, the rest of the video exists to reinforce and justify the click. Most successful TOS creatives follow a simple structure:
Hook (0–2s): product and use-case clearly on screen.
Clarify (2–8s): short, on-screen phrases or visuals that explain the main benefit (“No-Spill Travel Mug”, “Relieves Neck Tension”, “Keeps Food Fresh 2x Longer”).
Proof (8–20s): close-ups, demonstrations, social proof cues (ratings, “over 10,000 customers”, certifications – within policy).
Nudge (last few seconds): a calm, clear prompt to act (“Choose your colour”, “Perfect for home and office”, “Shop now on Amazon”).
Audio is a bonus, not a requirement. Many shoppers will never turn the sound on, so your story has to work safely silent. That’s why clear text overlays and strong visual storytelling matter more than scripted voiceovers.
Designing specifically for Amazon (and mobile)
TOS video is not just a cut-down version of a brand film. It’s a performance asset tailored to Amazon’s environment:
Length: 15–30 seconds is usually enough. Longer videos risk drop-off: shorter ones may feel rushed.
Aspect ratio: follow Amazon specs and always preview on a real phone. Make sure no key text or product detail is cropped by safe areas or UI elements.
Text overlays: use large, bold, sans-serif fonts with high contrast. Assume the viewer is on a small screen in bad light.
Background: keep it simple. The product and the benefit should be the focal points, not the set design.
Think of your video as a moving main image with context, not as a TV ad. If it doesn’t make the product clearer than a static photo would, it’s not doing its job.
A/B testing: the only way to find your real “winner”
Even seasoned creative teams guess wrong. What feels powerful on a storyboard doesn’t always translate into higher CTR and conversion. That’s why testing is not a luxury; it’s the core of making TOS video work.
There are three high-impact elements to test first: the opening frame or first shot (product close-up vs lifestyle; static vs in-motion), the first on-screen message (outcome-based vs feature-based), and visual style (bright, high-contrast look vs softer, more neutral treatment).
Keep each test focused. Change one major thing at a time and run both versions under similar conditions (same keywords, similar bids and budgets, similar timeframe). Then compare CTR, view-through rate or average watch time, and conversion rate.
Once you find a better-performing variant, treat that as your new baseline – then iterate again. The aim is not to chase perfection; it’s to keep climbing a staircase of small, measured improvements.
When to refresh – and when not to overthink
A good TOS video can work for months, but nothing is permanent on Amazon. Competitors upgrade their creatives, seasons change, and shopper expectations shift. If you see CTR and conversion gradually softening over time, it’s a sign your video has gone from “fresh” to “fine” – and fine rarely wins.
That doesn’t mean you should rebuild from scratch every month. Often, a new hook shot, a tweaked message, or refreshed visuals that reflect a season or use-case are enough to restore performance.
The key is to treat TOS video as a living part of your ad system, not as a set-and-forget asset. Because at the top of search, you are paying for every millisecond of a shopper’s attention. The first two seconds decide whether that money buys you a chance – or just a passing blur in someone’s scroll.

