Meta Ad Library

How the Meta Ad Library works

The Meta Ad Library is a transparency tool. Meta first launched it in May 2018, focused on ads about politics or social issues 1. On March 28, 2019, Meta updated it to include all active ads any Page is running, not only political ones, in one place 2. The same update let people search by Page as well as by keyword 2.

Two kinds of records sit in the library.

• Commercial ads appear while they are active. Each record shows the ad creative, the Page that runs it, when it started and the platforms it runs on.
• Political and issue ads get a fuller record. Meta says the library shows who saw these ads, their spend and impressions, and keeps them for seven years 2.

European rules added a third layer. On August 22, 2023, Meta announced that ads targeting people in the EU would be displayed and archived with the dates they ran, the targeting parameters used and who was served 3. Meta said these ads would be stored in the public library for a year 3.

Page details sit alongside the ads. Since 2019, Page transparency information includes the date a Page was created, previous Page merges and name changes 2. Those details help tell a long-running brand from a new or renamed Page.

Sales and marketing teams use the library as a source of ad transparency data. A brand running many long-lived ads is spending on growth and has a working funnel to sell into. The ad copy also gives a rep a specific first line.

The library describes Pages rather than companies. RichAPI's Meta playbook points out that one brand can run several Pages, such as regional and sub-brand Pages, and that the link on an ad is often a store platform subdomain or a link shortener rather than the brand's own site 4. Turning ads into a list of companies takes a bridge step, followed by normal data enrichment to reach firmographics and contacts.

What the Meta Ad Library does and does not show

Several limits shape what the library can tell a seller.

Spend is mostly hidden. RichAPI's playbook notes that spend is not disclosed outside the political and issue ad regime, so a figure for a commercial advertiser is an estimate 4. Use run length and ad volume as proxies instead.

Run length is the best commercial signal. The playbook argues that an ad running continuously for four months is an ad that works, while a Page with thirty ads all started last week is testing or being run by an agency 4. Rank Pages by their earliest active start date.

Pages are not companies. Brands run regional, seasonal and sub-brand Pages, and no dedupe rule catches every variant 4. Review the grouped list by eye before paying for contacts.

Agencies blur the buyer. Performance agencies run ads from the client's Page, so the Page looks like the brand while creative decisions sit elsewhere 4. That suits sellers who target brands and misleads those who target agencies.

Small brands dead-end. A small direct-to-consumer brand may have a Page and a store but no company profile worth matching 4. Expect a share of Pages to drop out before the contact step, which is a cost to plan for when comparing this with technographics or other fit filters.

Political ads are a separate regime. Their fields and retention differ, so filter them out of a commercial prospecting run 4.

Example of turning Meta Ad Library results into an account list

A team selling creative testing software wants direct-to-consumer brands with ads that have been running for months. It follows RichAPI's Meta recipe, which starts from an Ad Library search URL and ends with verified contacts 4.

1. meta_ads_library_scraper_sync     Ad Library search URL, limit set on purpose
2. group by pageName                 keep the earliest startDate per Page
3. resolve each Page to a website    use ctaDomain when it is a real domain
4. linkedin_company_search           match the website to a company
5. people, then email_finder         finder misses return billed false

The scrape is the first cost. meta_ads_library_scraper_sync takes a meta_ad_url and a limit of up to 100, and returns records with fields such as adArchiveID, adStatus, adText, ctaDomain, pageName, publisherPlatforms and startDate 4. It bills 1 credit per result 4. In the playbook's worked run, a limit of 100 returns 100 ads for 100 credits, and those ads collapse to 38 Pages 4.

Grouping comes next. The team groups by pageName and keeps the earliest startDate for each Page, because a long-running ad points to a working funnel 4. A Page whose ads all started last week goes to the bottom of the list.

The bridge step resolves Pages to companies. The team uses ctaDomain when it is a real corporate domain 4. A store platform subdomain tells the team the platform and little else, so those Pages need a separate website lookup 4. The resolved domains then go to linkedin_company_search, which bills 0.1 credits per result 4.

The playbook also warns about scope. At 1 credit per result the scrape costs a third of the Google equivalent, which makes wide pulls affordable but moves the money risk to the middle of the chain 4. A hundred ads from one busy advertiser can collapse to three Pages 4.

Each rep then quotes a running ad in the first line. The ad is public and dated, so the buyer can see exactly what the rep saw.

Meta Ad Library vs Google Ads Transparency Center

Both are public ad disclosure tools that sellers mine for account lists. They differ in what they identify and which advertisers they cover.