Skip to main content
The competitive researcher is where most good creative starts. Superscale finds the brands you compete with and pulls the ads they’re currently running, so you can see what’s working in your niche instead of guessing at it. Competitor research goes wider than your direct rivals. The point is to find ads that solved the same marketing problem you’re trying to solve.

​
How it works

The agent finds your competitors

Working from your brand and website, Superscale discovers competitors through web search and reasoning, then ranks them by how directly they compete. You can add or remove any of them.

It pulls their live ads

Superscale surfaces the ads each competitor is running across ad libraries. You see real, in-market creative, not a stale swipe file.

You spot what's working

Browse and filter the feed to find hooks, formats, offers, and visual patterns that keep reappearing. Those repeats are the patterns worth testing.

​
Choose the right competitors

Direct competitors

Brands selling the same thing to the same buyer. Start here when the market is clear.

Same-league competitors

Brands with similar size, budget, visual quality, and channel maturity. These are often more useful than giants.

Inspiration sources

Brands outside your category that solve the same objection, trust problem, or value proposition.
If the agent finds competitors that are too broad, too famous, too local, or simply wrong, correct the list. Add the brands you already know, remove irrelevant ones, and tell the agent what makes a competitor relevant for this task.
“Same-league” is often the best filter. A tiny local brand can learn from a category giant, but the sharper lessons usually come from brands with similar resources that are still running strong creative.

​
How Superscale scores winning ads

Superscale does more than show you what’s running. It scores every ad it surfaces with a winning-ad score from 0 to 100 that predicts how well that ad is performing for the brand running it, then puts the proven performers first. Public ad libraries never expose private ROAS, so the score reads the next best thing: how advertisers behave. The logic is simple. Advertisers cut what doesn’t work and pour budget into what does, so the ads a brand keeps alive and keeps multiplying are almost always its winners. Three public signals turn that behavior into a number:
  • How long the ad has been running. Losers get cut fast, so a long run is the strongest sign of sustained performance.
  • How many variations of the same concept the advertiser is running at once. When a brand floods its account with ten cuts of one idea, it has found a winner and is scaling it.
  • How widely the ad has reached.
Each ad is graded against that advertiser’s own baseline rather than one global bar, so a scrappy DTC brand and a category giant are judged fairly instead of by raw budget. Scores above the proven-performer mark are the ones worth studying and replicating, and weaker signals filter out. The score recomputes continuously as ads come and go.
82

Proven performer

Running 134 days · 9 variations of this concept · wide reach

​
It works on your own brand too

The same score runs on your own ads in the ad library, and that is where it tends to surprise people. In a demo, before anything is connected, Superscale can pull a brand’s own account and point straight at its top performers, the exact ads the team already knows are working. Watching the model find your winners from the outside, with no access to your numbers, is usually the moment the research clicks.
Product walkthrough: the research feed, scored and ranked by winning ads
Treat the score as a strong hypothesis, not proof. It is a proxy for performance inferred from public behavior, not a guarantee the same ad will win for your brand. If you have connected Meta, lean on your real account data for your own ads.

​
What you can do in the research feed

Use the feed as a working surface, not just a gallery.
ActionUse it for
FilterNarrow by media type, platform, status, language, category, time range, favorites, and sort order.
FavoriteKeep useful ads close for later analysis or recreation.
Save to DriveStore an ad as a reusable reference asset.
View originalOpen the source ad-library entry where available.
DownloadExport the creative asset when the source and file type allow it.
ReplicateRecreate the structure for your brand when the ad is an image creative that supports cloning.

​
Turn research into creative

Analyze why it works

Break a competitor ad down into its hook, structure, and signals.

Recreate it for your brand

Take a proven ad and make your own version of it.
The strongest prompt is specific about what to keep and what to change:
“Use this competitor ad’s structure: problem hook, fast proof, product demo, direct CTA. Replace the product, claims, visuals, and voice with our own brand context. Keep only the pattern.”

​
Common research questions

Copy the structure, not the brand. Learn from the hook, proof, pacing, framing, and offer, then translate those patterns into your own product and audience.
Add same-league or adjacent brands. Look for companies with similar buyer objections, price points, or trust problems even if they sell something different.
Manually add your list and tell the agent what to ignore: geography, enterprise brands, low-quality pages, marketplaces, or unrelated use cases.
Superscale starts from available ad-library data. If you have your own swipe file or past winners, add them as references so the agent can learn from them too.
Save the winners you find. A proven ad is the best brief there is, and it becomes a reference the agent can build on.
Last modified on June 4, 2026