Every day, your competitors spend money testing ads. The ones that stay live for several weeks have been validated by the market: not by an opinion, by a budget. Cloning competitor ads means using that already-paid-for testing as the starting point for your own, instead of going back to a blank page for every creative.
This page walks through the whole manual method, what it gets you, where it gets stuck, and how we automated it.
What “cloning” means, and what it doesn't
Let's clear up the ambiguity right away, because it wastes a lot of people's time.
Copying a competitor's visual (reusing their photo, their logo, their text word for word) is infringement. It's illegal, Meta penalizes it, and above all it doesn't work: the visual shows their product, not yours.
Cloning the structure is something else. An ad that performs rests on a repeatable mechanism: an angle of attack, a promise, a hook format, a visual hierarchy, a proof point. Those elements can't be protected. They're ideas, and nobody owns an idea. It's what every good media buyer has always done.
So the useful question isn't “which image do I grab”, but “why has this ad been holding up for six weeks, and how do I recreate that mechanism with my product”.
The manual method, in five steps
It's free and it works. If you do nothing else after reading this page, at least do this.
1. Open the Meta Ad Library
Go to facebook.com/ads/library. It's a public database where Meta is legally required to disclose every ad running on Facebook and Instagram. No account needed, no limit on how much you can browse.
Select the country, choose the “All ads” category, then search for either a specific advertiser name or a keyword related to your product.
2. Identify the right advertisers
Search by product keyword rather than by brand: you'll discover competitors you didn't know about. A brand that advertises a lot, regularly, and has done so for a long time is a good case study: it has a budget, so it probably has a return.
Keep a list. Five to ten serious advertisers in your niche is more than enough.
3. Spot the creatives that last
This is the step most people skip, and it's the only one that really matters.
The library shows no performance data: no CTR, no cost, no sales. It's a transparency tool, not a dashboard. The only usable signal is the date the ad started running.
An ad that has been live for eight weeks is almost certainly profitable. Nobody pays to run a creative that loses money. A creative rolled out in ten active variations is an even safer bet: it means the advertiser has decided to invest in it.
Ignore anything less than two weeks old. Those are tests, and you don't know yet whether they'll make it.
4. Break down the angle
For each creative that lasts, note four things:
- The angle, what the ad goes after: price, fear, result, novelty, comparison, testimonial?
- The promise: what specific benefit is announced, and in how much time?
- The form: a chat screenshot, a before/after, the product in use, full-frame text, a fake organic post?
- The proof: a number, a customer review, a badge, a visual demonstration?
After twenty creatives, patterns emerge. You'll see that three or four angles come up again and again in your niche. Those are the ones that work.
5. Recreate the mechanism with your product
You don't copy anything. You take the angle that comes up most often and apply it to your product, with your selling points and your proof, in your brand style. Then you test.
Where the manual method gets stuck
It doesn't get stuck on quality: it gets stuck on volume and consistency.
The library has no memory. You can't know that a creative went live six weeks ago and was taken down the day before yesterday. You see a snapshot, not a history.
The sorting is manual and slow. Allow one to two hours to explore a niche properly. And you have to start over every week, because competitors launch new creatives all the time.
Going from analysis to production is the real friction. You've identified a good angle. You still have to write the hook, produce the visual, adapt it to your product and make five variations to test. That's where most brands give up, not for lack of ideas, for lack of time.
Who this approach isn't right for
Better to be clear now: it'll save you some disappointment.
If you're launching a brand in a market with no identifiable competitor, there's nothing to clone: the library will be empty for your niche, and half the value of the method disappears. You're better off starting with the “customer reviews” angle.
If you're after original art direction, starting from existing ads will, by design, keep you inside your market's visual vocabulary. That's effective for converting, less so for standing out.
If your main channel is video, no worries: static or video, you clone the structure of the ad and recreate it with your product.
How CreaFlow AI automates these five steps
| Step | By hand | With CreaFlow AI |
|---|---|---|
| Finding advertisers | Keyword search, over and over | Competitors tracked, new ads detected automatically |
| Creative history | Snapshot only | History kept, including ads that were taken down |
| Spotting what lasts | Reading dates by hand | Sorting by running time and by number of variations |
| Understanding the angle | Manual analysis, creative by creative | Angle, promise and format extracted automatically |
| Producing your version | Copywriting and design, creative by creative | Variations generated with your product and your brand DNA |
The principle is simple: you never generate from a blank page. The starting point is always an ad that is actually running in your niche, to which the tool applies what it knows about your brand: your products, your angles, your personas, and your own customers' vocabulary.
In practice: you give the link to your store, the tool analyzes your brand and your products, you choose which competitors to track, and you get back creatives that are ready to test, static and video alike.
In one sentence
The five steps described above take half a day a week if you do them seriously. CreaFlow AI runs them back to back for you: the monitoring runs continuously, the ads that last surface on their own, their mechanics are analyzed, and your variations come out with your product and your brand DNA. You keep the only decision that matters, which ones go into testing.
The Chrome extension also lets you capture an entire brand while you browse, and ad angles let you vary what you test. The free plan lets you try it all with no credit card.
Page written in August 2026. The Meta Ad Library is a service of Meta Platforms, publicly accessible and free of charge; the brands mentioned belong to their respective owners.