How to Read a Competitor’s Ads in the Meta Ad Library

Meta Ad Library research shows you every ad a competitor is running on Facebook and Instagram today: the creative, the copy, the date each ad started and the page most of them send people to. For US product ads it shows no spend, no impression counts and no results, so the work is reading the signals the Library does give you and being clear about what it cannot tell you.

This guide is the method behind our own Ad Library sweep. On October 2, 2026 we ran 12 keyword searches in the US Ad Library across supplements, pet, food and beverage, beauty and outdoor gear, read 461 ad cards from 187 advertisers, and then opened the full Page view of 31 brands. The numbers below come from that sweep, written up in full in our DTC Meta ads study. Genti Xharavina, who wrote this guide, spent years making paid-social creative for direct-to-consumer brands before NextAds.

What the Meta Ad Library shows, and what it does not show for US product ads

The Ad Library is Meta’s public, searchable record of the ads running across its apps, Facebook and Instagram included. Many people still call it the Facebook Ad Library. Meta’s help page says you can search every ad that is active today, and inactive ads only when they are about social issues, elections or politics. A live card in a US search carries these fields.

• The ad. The image or video, the text above it, the headline, the website and the button, as people see them.
• Start date. “Started running on” followed by a date. That is how you work out how long an ad has run.
• Platforms. Icons for the Meta apps the ad is set to run on, such as Facebook and Instagram. Compare them across a brand’s ads to spot the ones that run on one app only.
• The advertiser. The Facebook Page that runs the ad. A brand can run ads from more than one Page, and a partnership ad joins a second name to the brand’s with “with”.
• Copies and versions. “2 ads use this creative and text” means the same ad is running more than once. “This ad has multiple versions” usually means several images, videos or lines of text were loaded into one ad, and the card shows only one of them.
• The link. The page the button opens, and the most useful field on the card. We come back to it below.
• Library ID. A number for each ad, so you can note exactly which ad you looked at and find it again.

What you want to know	For a US product ad
Is it live, when did it start, where does the click go?	Yes, on almost every card. A few ads have no website link, and a few hide their details until you log in to confirm your age.
How much is the brand spending?	No. Spend ranges appear only on ads about social issues, elections or politics.
How many people saw it?	No number. Results are sorted by impressions, high to low, unless you change the sort.
Who is it aimed at?	No. Among product ads, only those delivered in the EU or associated territories show the locations targeted and reach by location, age and gender.
Is it selling?	No. Nothing in the Library measures sales, return on ad spend or cost per order.
What did they run last year?	Mostly no. Meta’s tools page says ads delivered in the EU or associated territories are archived for a year after their last impression. Otherwise, Meta’s help page lets you search ended ads only when they are about social issues, elections or politics.

Ads about social issues, elections or politics are the exception: they show who paid, a range of spend and reach, and stay in the Library for seven years. Meta’s Ad Library tools page also describes a separate search for branded content, the active posts with a paid partnership label, which shows a brand’s labeled creator posts.

If a competitor also advertises in the EU, set the country to an EU market and open See ad details on one of its ads. Under Transparency by location you will find the locations it targeted and its EU reach, broken down by location, age and gender.

Find every ad a brand runs, beyond the ones your keyword caught

A lot of people who set out to spy on competitor Facebook ads type a keyword and stop there. A keyword search returns ads that match your words, which makes it good for finding competitors and poor for sizing them. Of the 154 advertisers we kept from our keyword sweep (mostly product brands, though the automatic filter also kept a handful of retailers, marketplaces, an online grocer and link or landing-page services), 103 showed up with a single ad. When we opened the Page view of 31 smaller-looking brands, 25 of them with only one or two keyword hits, the median brand had at least 29 live ad cards.

Keyword results are long, too. A US search for “dog chews” showed about 18,000 results on October 2, 2026, a count that includes marketplaces and retailers, and we read the top 38. Here is how to get from a keyword to a brand’s full list.

1. Set the search. Country: United States. Ad category: All ads. In the Filters panel, leave Media type on all media types, or you will miss either the statics or the videos. The Platform filter narrows the results by Meta app.
2. Open the brand’s Page view. Type the brand’s name in the search box and choose its Page from the suggestions. That opens every ad the Page runs, whatever words the ads use.
3. Set Active status to Active ads. These are the ads the brand pays for today. With the status on Active and inactive, 6 of our 31 pages showed some ads with an end date and the Library’s count ran far above the live ads we loaded. We did not load the rest, so do not read that list as a complete history. Meta’s tools page says ads delivered in the EU or associated territories are archived for a year after their last impression, which may explain some of the ended ones. The Library does not say why an ended ad stays visible.
4. Switch Sort by to Most recent when you count launches. The default, Impressions: high to low, puts long-running ads first. Most recent brings the newest ads to the top.
5. Scroll until the cards stop loading. Cards load in batches. We loaded only the first batch, up to 40 cards. On 21 of the 31 pages every card we loaded was live and the Library’s own total was higher, and our live counts on those pages are minimums. In our checks the Library’s total counted every copy of an ad (and ended ads when the status was Active and inactive), so it can run above the number of cards even when the page has fully loaded.
6. Count by the link’s website. Grouping ads by the domain in the link catches most ads from a brand’s second Page and most partnership ads, where a creator’s or publisher’s name is joined to the brand’s with “with”. A few partnership ads send the click to the creator’s Instagram profile or to a marketplace, so read the advertiser line too. Grouping this way turned 461 keyword cards into 187 advertisers, and 17 of the 154 advertisers we kept had partnership ads in the sample.

Five signals an ad is earning its budget

None of these is a Meta metric, and Meta does not say that run time, copies or result order mean an ad is working. They are rules of thumb, each has an innocent explanation, and they are only worth trusting in combination.

• Run time. An ad left on for three months has either been paying its way or been forgotten. Among the 336 dated ads from the 154 advertisers we kept, the median had run 130.5 days, and 235 had run 90 days or more.
• Duplicates. When several ads use the same creative and text, the brand has put that ad in more than one place. Copies of an ad that has run for months say more than copies of one launched last week.
• Fresh launches. With Sort by on Most recent, count the ads that started in the last 30 days. A steady flow means the brand is still testing. A page that seems to have none may just have more cards than you have scrolled through.
• Repeated hooks. Group the ads by their first line and the opening of each video. When a brand keeps writing new ads around the same opening, that line is doing the selling.
• Result order. Unless you change Sort by, results run from most impressions to fewest. A high position means more impressions than the ads below it, with no number shown, so read it as a rough sign of reach.

Read them together

An ad that has run for six months, sits near the top, has copies and shares its opening line with the brand’s newest launches is the strongest case the Library can make. An old ad with no copies and no newer ad reusing its opening may simply be one nobody has switched off. Ten ads with one opening line are one idea tested ten ways, so count ideas as well as ads.

Our own numbers lean old. We kept the default sort and read the top 35 to 40 ads of each search, so long-running ads crowd the top and new launches are under-counted. Your searches will do the same unless you switch to Most recent.

Follow the click: what the landing page tells you

The link is the hardest fact on the card. The page a brand pays to send people to tells you what it is selling, to whom, and how much a first-time buyer has to read before ordering. Sort each link into one of these.

• Homepage. The ad asks a new visitor to find the product on their own. It is the easiest gap to spot in a competitor’s account, and in your own.
• Product page. One product and one offer. Read the title, the first images and everything above the add-to-cart button, because that is what the ad has promised.
• Dedicated landing page. Links with /pages/ or /lp/ in the path, or subdomains such as try. or get., often lead to a page built for paid traffic. Open it before you count it, since a /pages/ link can be an ordinary content page. If the ad’s first line and the page headline match, you have found the angle the brand is paying to test.
• Advertorial or listicle. A story or a list of reasons before the product, used for products that need explaining before anyone will buy.
• Discount link. A link that applies a code and then redirects. Note where it lands and whether the code shows on the page.

Then read the page the way a new buyer would, on a phone: the subscribe option, the bundles, the review count and how long it takes to load. The ad is only the front of the funnel, and the store shows how much of the rest the brand has built. Most of the stores in our sweep ran on Shopify, and most had an email tool and a review app installed. The counts are in the DTC Meta ads study. To test the same idea on your own store, see landing page testing.

Check Google’s Ads Transparency Center too

Google keeps its own public record. The Ads Transparency Center lets you search by advertiser name or website and filter by region and date, according to Google’s help page on advertiser verification. Google’s launch post describes it as a searchable hub of ads from verified advertisers across Search, YouTube and Display.

It is worth the extra minute. 62 of the 72 brands on our shortlist had at least one ad there. A Google Ads tag showed in the homepage code of 51 of the 72, yet 13 brands running Google ads had no tag we could see, so look the brand up instead of reading its code.

1. Search by website. We looked up all 72 brands by domain, which avoids guessing at the legal name behind an advertiser account.
2. Set the region to United States. Write down the date you looked and the date range the Center applied, because the count changes with both.
3. Put Google next to Meta. Text ads show how the brand describes itself to people already searching. Image and video ads show which products it backs. Look for the offer that appears on both.

Three limits apply. A count of ads says nothing about spend, and 17 of our counts were Google’s own rounded figures. We kept the Center’s default date range without recording it, so treat our counts as a snapshot. And our 72 were a shortlist picked from the sweep, not a random sample. For Google search ads run alongside Meta, see Google Ads management.

What you still cannot know

Neither the Ad Library nor Google’s Ads Transparency Center answers these. Paid tools that save copies of ads as they find them can fill in part of a brand’s ad history, and any spend figure they show is an estimate. Unless an answer comes from the brand’s own account, treat it as a guess.

• How much a brand spends on any ad, campaign or month.
• How many people in the US saw an ad, and who they were.
• Whether a long-running ad is profitable or simply left on.
• Which version of a multiple-version ad gets most of the delivery.
• What the brand pays to win a customer, and what that customer is worth over a year.
• What happens after the click: the emails, texts and retargeting that turn a first order into a third.
• Most of what the brand ran before today, because ended product ads drop out of view.

Use the Library for ideas, gaps and timing, and your own account data for decisions. Inside your account, the numbers depend on tracking that reports each order back to Meta. Our guide to the Meta Pixel and Conversions API on Shopify covers that side.

If you would like this done for your brand, our free Meta ads audit applies the same reading. It is built from public information only: your live ads in the Meta Ad Library, the pages they send people to and what your Pixel shows from the browser. You do not need to give us access to your ad account. It ends with three ads or angles we would test first. A partner replies within one business day, the written teardown follows by email, and we never publish a teardown without the brand’s written permission. When you want the ads run as well, see Meta ads for ecommerce brands.

Ad Library questions, answered

Can you see how much a competitor spends on Facebook ads?

No, not for product ads in the US. Meta shows spend ranges only for ads about social issues, elections or politics, so any spend figure for a competitor’s product ads is an estimate, whoever quotes it.

Is the Meta Ad Library the same as the Facebook Ad Library?

Yes. It is the same tool under Meta’s name, and it covers ads across Meta’s apps, Facebook and Instagram included.

Do I need a Facebook account to use it?

No. You can search and read ads without logging in, though a few ads hide their content until you log in to confirm your age.

Why can’t I find all of a competitor’s ads?

A keyword search only returns ads that match your words. Start from the brand’s Page view, scroll until no more cards load, and group the ads by the link’s domain to catch other Pages and partnership ads. An ad that has ended is usually gone from view.

How long does an ad need to run to count as a winner?

Meta publishes no threshold. A workable rule of thumb is to start paying attention at 90 days and look hardest past 180, especially when the ad has copies and newer ads reuse its opening. In our October 2026 sample, 51 of the 153 advertisers with a dated matching ad had one running past 180 days. Run time alone can also mean nobody switched the ad off.

Can I copy a competitor’s ad?

You can’t reuse the ad itself. Meta’s Advertising Standards bar ads that violate a third party’s intellectual property rights, copyright and trademark included. The footage, photos, music, wording and customer quotes belong to the brand and its customers. What you can take is the lesson: the angle, the type of page, the offer structure and the timing, rebuilt with your own product and your own proof. That is where our ad creative work starts.

How often should I check a competitor’s ads?

Monthly is enough for most brands, and weekly in the run-up to a big sale. The Library only shows what is live on the day, so each time write down the date, the live count and the start dates. The changes between two visits tell you more than either visit alone.