If you want to A/B test your Amazon video ads properly then stop putting them in
By Christophermrawlings · February 17, 2026 · Curated by George's Blog
If you want to A/B test your Amazon video ads properly then stop putting them in separate campaigns. Do that and you’re just gambling.
When you split campaigns Amazon changes a bunch of stuff behind the scenes like where it shows your ad + how often it shows it + how fast it spends. So this way you don’t know whether the video won or Amazon was just giving it better placements.
Instead run multiple Sponsored Brand Video ads inside the SAME campaign = same keywords, same bids, same budget, same placements - only the video changes.
This works because you’re getting a real comparison between them. There’s no placement bias where top of search gets all the love. And when a video wins you can scale it way more confidently instead of just guessing.
What we test with one change at a time:
- First 2-3 secs of the video (basically the hook)
- Benefits vs lifestyle vs problem/solution angles
- Product only vs the product being used
- Quick Canva job vs UGC creative
We only care about 2 numbers - thats the CTR because it shows us who stops and clicks, and the ACoS which shows how efficient the ad is.
Doing this got us $8.9K at 18% ACoS which beat $4.2K at 20% ACoS, all just from a creative swap.
Comment "SBV" and i will send you our Sponsored Brand Video framework.