Meta Ad Library
The Meta Ad Library is a public, searchable collection of ads running across Meta’s apps, including Facebook and Instagram. You can search by advertiser name or keyword, filter by country and see the creative for each active ad. No account is needed to browse.
Start with brands in your category and the brands your customers also buy from. Open their active ads and look at the formats they use most. When one brand is running several versions of the same idea with different hooks or products, that idea is probably worth studying.
TikTok Creative Center
TikTok Creative Center is TikTok’s own hub for advertisers. It includes a library of top ads you can filter by region, industry and objective, along with trends, popular songs and creator insights. It is useful for seeing which formats are common on TikTok right now, which may differ from what runs on Meta.
Save notes on the ads that stand out: the hook, the format and how quickly the product appears. Do not download other brands’ videos for reuse.
How to tell if an ad is “winning”
Public tools rarely show performance numbers for other brands’ ads, so be careful about claims like “this ad made millions.” Instead, look for signals you can actually observe:
- How long it has been running. Advertisers tend to stop paying for ads that do not work, so an ad that has run for a long time is a reasonable sign of performance. The Meta Ad Library shows when an ad started running.
- How many versions exist. Several variations of one concept suggest the advertiser is investing in it.
- Whether it reappears. An ad that keeps coming back in a new season or with a new offer has probably earned its place.
- Your own data. Your own ads, with your own results, are the most reliable references you have.
These are signals, not proof. Treat them as reasons to study an ad, not as guarantees it will work for you.
Clone the structure, not the footage
Once you have a shortlist, write each ad out shot by shot: hook, shot order, pacing, product reveal, captions and call to action. That breakdown is the part you can reuse. Then rebuild it with footage you own or have licensed, your product images, logo and copy.
Read is it legal to copy a competitor’s ad for where the line sits, and clone viral video ads for the full workflow in Clone Cut. For Meta-specific formats, see clone winning Meta ads with AI.
A simple research routine
- Each week, check the Meta Ad Library for a few brands in your category.
- Check TikTok Creative Center for top ads in your industry and region.
- Log any ad that has run for a long time or has several versions.
- Write a breakdown for the two or three most promising.
- Rebuild one as a test with your own assets, changing one variable at a time.
Make only claims you can prove, disclose AI-generated content where the platform or law requires it, and never publish fake testimonials or endorsements.
What to write down for each ad
A screenshot is not enough. For each ad worth studying, record the same fields so you can compare ads later:
| Field | What to note |
|---|---|
| Advertiser and date | Who runs it, and when it started running if shown |
| Format | Testimonial, unboxing, problem-solution, listicle or another format |
| Hook | The first line and the first shot |
| Product reveal | The second the product first appears |
| Pacing | Roughly how many cuts and how long the ad runs |
| Offer and call to action | What the viewer is asked to do |
| Versions | How many variations you can see |
Over a few weeks this becomes a swipe file of structures, not footage, and patterns start to stand out.
Signals that can mislead
- Engagement on organic posts. A post with many likes is not necessarily a profitable ad.
- A single snapshot. One day’s view of a library tells you little; check again over time.
- Category differences. A structure that suits a low-priced impulse product may not suit an expensive considered purchase.
- Claims in third-party roundups. Lists of “top ads” rarely show where their numbers come from. Prefer what you can observe yourself.
Turning research into a test
Pick one structure from your notes, write a brief that keeps it and changes only your product and copy, and run it against your current best ad. If it wins, test new hooks on top of it. If it loses, note why you think it did and try the next structure. The video ad hooks guide has opening lines to test, and the video ad templates give ready-made structures by format.
Research your own account first
Your own ads are the best references you have, because you know exactly how they performed and you own the footage. Sort your past ads by the result that matters to you and write a breakdown for the top few. Look for patterns: the formats that keep winning, the hooks that keep working and the products that respond to a particular structure. Then compare those patterns with what you saw in the public libraries.
When an ad of yours has worked, it is the safest kind of reference to clone. You can keep the people and the structure and change the product, the background, the logo or the copy without any question about rights.
Ethics of competitive research
Looking at public ad libraries is normal and expected; that is what they are for. Using what you find responsibly means learning from structure and ideas, not lifting footage, scripts, music or people. It also means not imitating a competitor so closely that customers could confuse the two brands.
How often to repeat the research
Ad libraries change every day, so treat research as a routine rather than a one-off project. A short weekly check is usually enough to spot new formats in your category. Once a month, review your swipe file, drop structures that no longer appear and pick the next one to test. Keep your notes in one shared place so your whole team works from the same references.
Frequently asked questions
Is the Meta Ad Library free?
Yes. It is publicly available and can be browsed without logging in.
Can I see how much a competitor spent on an ad?
Generally not for ordinary commercial ads. Spend data is shown for some categories, such as ads about social issues, elections or politics. Use running time and the number of variations as public signals instead.
Can I download ads from these libraries to reuse?
Use them for research only. The footage belongs to the advertiser and its creators. Rebuild the structure with your own or licensed footage.