How to Find Winning Products for Ads | Polaris
How to identify which products are worth building ad creative around, and why testing creative volume is part of the discovery process itself.
Finding winning products for ads means identifying which items in your catalog (or which items you're considering sourcing) have a clear enough hook, visual appeal, or problem-solution story that they can carry a scroll-stopping ad — and then actually testing creative against them, since a product's real potential often only shows up once you see how an audience responds to ads for it, not before.
What makes a product a good candidate for ads in the first place?
Products that photograph or demo well, solve a clear problem, or have an obvious "before and after" tend to translate into stronger creative than products that are hard to explain visually or whose benefit is abstract. A product doesn't need to be new or novel to work — it needs an angle that's easy to communicate in a few seconds of video or a single image, since that's what the ad has to do in the opening moments.
Can I tell if a product will win before running any ads?
Not with certainty — signals like how well a product demos, whether it has a clear pain point it solves, and how it compares visually to competitors in the space are useful filters, but the real test is whether an audience actually stops and engages with an ad for it. Treat early signals as a way to prioritize which products to test first, not as a guarantee, since actual performance depends on execution, audience, and timing as much as the product itself.
How many products should I test creative for at once?
There's no fixed number — it depends on your catalog size, budget, and how much creative production you can realistically support. The more products you can put a real ad in front of an audience for, the more chances you have of finding one that resonates, but each one needs at least a genuine creative attempt (a real hook, not a throwaway) to be a fair test.
Why does creative production speed matter for product testing?
The bottleneck in finding winning products is usually not the products themselves — it's how quickly you can produce a real ad for each candidate to actually test it. If producing one ad takes days, testing ten product candidates becomes a multi-week project, and by the time you learn which ones work, timing and trends may have shifted. Faster creative production directly means faster product discovery.
How does Polaris fit into product testing?
Polaris connects to Claude as an MCP tool: you attach a product photo or paste a script and ask Claude in plain chat to generate the ad, picking settings from tappable option cards, and a live panel shows the finished ad rendering in the conversation in about 20 to 180 seconds. Because it builds ads around your actual product image rather than a stock-avatar library, and batches render in parallel and download as a ZIP, you can put together a real first-look ad for several product candidates in one session instead of one at a time — turning product research from a guessing game into something you can actually test with real creative, quickly. See the full workflow at AI UGC ads or read how the connection works in generating ads inside Claude via MCP.