How to Use the Meta Ad Library to Find Winning Ads

The Meta Ad Library is the most useful free tool in paid social, and most advertisers only scratch the surface. It shows every ad running on Facebook, Instagram, Messenger, Threads and the Audience Network, from every advertiser, in every country. If a competitor is spending money on Meta, their ads are in there.

This guide shows how we use it at adthief.ai to find ads that already sell, and how to read the few signals the library gives you so you can tell a winner from a test.

What the Meta Ad Library shows (and what it doesn’t)

Meta built the library for transparency, not for marketers, so it’s worth knowing its limits before you start.

What you can see for every ad:

The creative: image, video or carousel, with the post text, headline and button.
The date it started running, and whether it is still active.
The platforms it runs on (Facebook, Instagram, Messenger, Threads, Audience Network).
Whether the advertiser runs several versions of the same ad.

What you get extra for ads shown in the European Union: because of the EU’s Digital Services Act, ads delivered in the EU also show an estimated reach, a breakdown by age, gender and country, and they stay visible for a year after they stop.

What you never see for commercial ads: spend, clicks, conversions or return on ad spend. So you have to read performance from indirect signals, which is the useful part of this guide.

Step 1: Search the right way

Go to the Ad Library at facebook.com/ads/library and set three things before you type anything:

Country. Pick the market you sell in. An ad can run in one country and not another, and local winners are more useful than global ones.
Ad category. Choose “All ads”. The other option only covers ads about social issues, elections or politics.
Search type. Search by advertiser name to see one brand’s ads, or by keyword to see everything that mentions a product (for example “collagen”, “protein powder” or “dog food”).

Keyword search is the better way to discover new competitors. Advertiser search is better once you know who you are watching.

Step 2: Filter to what matters

After searching, use the filters to cut the noise:

Active status: active ads. A stopped ad tells you less, unless you are studying a brand’s history.
Media type. Look at images and videos separately; they win for different reasons.
Platform. If you only plan to run on Instagram, filter to Instagram.
Impressions by date. Set a date range to see what started in the last month, or what has been running since before it.

Step 3: Read the signals of a winning ad

Since the library hides performance, you read it from behaviour. Advertisers don’t keep paying for ads that lose money. Three signals do most of the work:

1. How long it has been running

The start date is the strongest signal you get. A new ad proves nothing yet. An ad that has been running for 30, 60 or 90 days has survived many rounds of budget decisions. We built our whole ranking around this: in our study of over 10,000 ads that ran for 60 days or more, the median ran for 77 days and one in ten passed 245 days. See what long-running Meta ads have in common.

2. How many versions the brand runs

When the library says “This ad has multiple versions”, or you see the same creative many times with different text, the brand is scaling it. Brands copy what works.

3. How far it reached (EU only)

For ads shown in the EU, the estimated reach tells you how much was spent behind an ad, roughly. Compare it with the brand’s other ads: a creative with ten times the reach of its siblings is where the budget went.

Step 4: Save, sort and compare

The Ad Library has no way to save ads, rank them or compare brands. This is where most people end up with a folder of screenshots. A simple system works better:

Keep a list of 5 to 15 competitors per market.
Check them weekly, and note every new ad and every ad that crosses 30 and 60 days.
Group the winners by angle: the problem they lead with, the promise, the proof, the offer.

If you would rather not do this by hand, adthief.ai does it automatically: it pulls the Ad Library for your countries every day, ranks every ad by how long it ran and how far it reached, and marks each brand’s best ads.

Step 5: Turn winners into your own ads

Copying an ad word for word is lazy and usually doesn’t work, because your brand, product and audience are different. What transfers is the structure:

The hook: what the first line or first second says.
The format: talking head, before and after, product demo, list, testimonial.
The layout of a static: where the product sits, how big the headline is, what proof is shown.
The offer: bundle, trial, guarantee, free shipping.

Take the structure, then fill it with your product, your proof and your customers’ own words. Our free Paid Ads Workbook walks through exactly that, from the brand brief to hooks, bodies and calls to action.

Common mistakes

Judging an ad by how good it looks. Ugly ads often run longest. Trust the dates, not your taste.
Only looking at big brands. A local competitor with a tight budget can’t afford to keep a loser running, so its long-running ads are strong signals.
Looking once. The library changes every day. The value is in watching it over time.
Ignoring the landing page. Click through. A winning ad often wins because of the page and the offer behind it.

The short version

Search your market by keyword, filter to active ads, sort mentally by start date, and study the ads that have run for 60 days or more. They are the ones the market has already voted for. Then borrow their structure, not their words.

Want this done for you every day? Try adthief.ai free, or talk to us about running your Meta ads.