Qu’est‑ce que la transparence publicitaire ? Bibliothèques, législations et comment les exploiter

Ad transparency used to be a courtroom topic. Now it is a research surface any media buyer, affiliate, analyst, or journalist can open in a browser and search. The shift took about five years, pushed along by election scandals, two big regulations, and a handful of platform launches. What you get today is a patchwork: a few genuinely good public archives, several mediocre ones, and one enormous category of advertising (native and programmatic display) that almost no official library touches.

This page is the hub for everything we publish on the subject. It covers what ad transparency actually means, where the data comes from, which laws forced it into existence, what the public libraries can and cannot show you, and how to turn the lot into usable competitive intelligence. Every section links out to a deeper guide.

For context on scale: the data we cite here comes from the OpenAdLibrary index, which as of June 2026 holds 589,036 captured native creatives from 25,933 advertisers across 42 ad networks, backed by 5.4 million ad observations and 926,259 landing-page captures. None of that lives in any platform's official archive. That is the gap this whole article is about.

What "ad transparency" actually means#

Ad transparency is the idea that the public should be able to see which ads are running, who paid for them, and where they appear, instead of that information sitting only inside the advertiser's account and the platform's servers. In practice it shows up as ad libraries: public, searchable archives of the ads a platform has served, usually tagged with the advertiser name and the dates each ad ran.

Split that definition in two and it gets clearer. The first half is disclosure at the point of view: the "Sponsored" or "Ad" label on a placement so a reader knows it is paid. The second half is durable, searchable archiving: a record that outlives the impression, so you can come back next month and ask "what was this brand running in March?" Most people mean the second when they say ad transparency. Regulators historically cared about the first. That is why the two keep getting tangled together. For a tighter, citable definition you can drop into a brief, see our Ad Transparency glossary entry.

Where the data comes from: three sources#

Transparency data does not arrive from one place. It comes from three structurally different sources, and the source decides how complete and how trustworthy the data is.

The first two overlap a lot, because most "regulator-mandated" repositories are just the platform's existing library with extra fields bolted on to satisfy the law. The third is a different animal. Instead of asking the seller what it served, an independent platform watches real ad slots on real publisher pages and records what actually rendered. That distinction matters most for native advertising, where there is no platform-run library to ask in the first place.

The cleanest test of any transparency source: does it show you what the seller chose to disclose, or what was actually served? When those two answers diverge, the second one is the one you want.

The laws that created modern ad transparency#

Ad libraries did not appear because platforms wanted to be helpful. They appeared, or expanded sharply, in response to regulation and the threat of it.

The EU Digital Services Act#

The most consequential rule is Article 39 of the EU's Digital Services Act (DSA). It requires every designated Very Large Online Platform to maintain a public ad repository covering all the ads it serves, not just political ones. The repository has to record what the ad promoted, who it was for, who paid for it where that differs from the beneficiary, the dates it ran, whether and how it was targeted, and roughly how many people in each member state it reached. (European Commission)

The DSA has teeth. The Commission fined the operator of X 45 million euros over a non-compliant advertising repository, and both TikTok and AliExpress agreed to binding commitments to bring their repositories up to standard. (European Commission) This is the single biggest reason ad transparency moved from a political-ads niche to an all-ads expectation.

The FTC and US disclosure rules#

The United States has no DSA equivalent, so there is no federal mandate for a universal ad archive. What the US has is a disclosure-at-the-source regime. The FTC's 2015 Enforcement Policy Statement on Deceptively Formatted Advertisements established that an ad which looks like editorial content is deceptive unless it is clearly identifiable as an ad. That is the legal root of every "Sponsored" tag you see on a native widget. It governs labeling, not archiving, which is exactly why the US leans on platform goodwill and independent tools for the archive half of the problem.

The practical upshot: in the EU, transparency is a legal obligation with audits behind it. In the US, it is a labeling rule plus a market of voluntary and independent solutions.

The major public ad libraries#

Three platform libraries do most of the heavy lifting in the official world. Know their exact scope and their exact limits, and you will stop over‑trusting them.

Google Ads Transparency Center#

Google launched the Google Ads Transparency Center on March 29, 2023, partly under DSA pressure and partly to answer the long‑standing critique that, unlike Meta, Google had no durable ad archive. It is a searchable hub of ads served on behalf of verified advertisers, indexing Search, Shopping, Display, YouTube, Gmail, Maps, and the Play Store. You can filter by advertiser, country, date range, and format, and Google has started showing the legal payer name on ad cards rather than just the marketing brand. (Google)

The catch is in the word verified. If an advertiser has not completed Google's identity verification, their ads do not appear. So absence from the archive is not proof an advertiser is inactive.

Meta Ad Library#

Meta's was the first major library and is still the deepest for social. It covers Facebook and Instagram, with full creative for active ads and, crucially, spend, reach, and funding data for ads about social issues, elections, and politics, which Meta keeps for seven years. Political‑issue ads have been tracked across more than 120 countries since 2020. (Meta) One wrinkle worth knowing: after the EU's political‑adverting rules tightened, Meta and Google stopped offering political ad services in the EU, so the "issues, elections, or politics" filter no longer returns results for EU nations even though general ad metadata stays available. (TechPolicy.Press)

TikTok and the rest#

TikTok runs a commercial content library and, under DSA pressure, committed in December 2025 to binding upgrades to its ad repository. (European Commission) Past these three, coverage thins out fast. We break down what each tool genuinely surfaces, and what it quietly omits, in What Is an Ad Transparency Tool? Open Ad Libraries Explained.

The big blind spot: native and programmatic ads#

Here is the gap that catches most practitioners off guard. The official libraries cover the platforms that own the auction: Google's surfaces, Meta's apps, TikTok's feed. They do not cover the open web's native and content‑recommendation ecosystem, where Taboola, Outbrain, MGID, Revcontent, Teads, MediaGo, Yahoo, and MSN deliver the "Around the Web" and "Recommended For You" widgets you see under articles.

There is no Taboola Ad Transparency Center. No public, searchable Outbrain archive. For years this entire category, which carries a huge volume of direct‑response and affiliate advertising, had no library at all. That is the story we tell in What Is a Native Ad Library and Why One Didn't Exist Until Now. The technical reason is that native ads are not served from a single queryable index. They are stitched into thousands of publisher pages in real time, personalized, and rotated constantly. To archive them you have to observe the live web, which is much harder than querying an API the platform already runs.

What that ecosystem looks like, once you actually capture it, is this:

These are real, sometimes shameless native ads. That is the point: this is the half of the advertising world the official archives skip. Finance is the single largest vertical in our index at 17,232 creatives, followed by insurance (15,629) and health (14,895). Taboola alone accounts for 157,727 of the creatives we have captured, Outbrain another 84,252, and MGID 49,689. None of those numbers exist in any seller‑run library.

This is the gap OpenAdLibrary was built to close. It captures live public native ads directly from publisher pages, stores the real creative image at full quality, classifies the ad‑tech supply chain behind each placement, and follows the click through to the advertiser's landing page (without clicking live ads). That gives the native ecosystem the thing it never had: a durable, searchable, independently verified record.

From transparency to intelligence#

Seeing an ad is the floor, not the ceiling. A raw library tells you an ad exists. It rarely tells you the four things that actually drive decisions:

1. Who is really behind it. The visible brand on a native ad is frequently an arbitrage or media‑buying layer, not the end advertiser. The only reliable way to find the real one is to follow the click to the landing page or pre‑lander, which the official libraries do not do.
2. What network delivered it. The same creative can run across multiple networks, and knowing which one matters for sourcing and for reading the competitive picture. We walk through the detective work in How to Identify the Ad Network Behind Any Ad.
3. How long it has run and how widely it has spread. Longevity and spread are the closest free proxies for performance you can get without the advertiser's dashboard. Industry lore says a creative still live after 90 days is almost certainly a winner, and that lore is broadly sound: nobody pays to keep a loser running. Treat that 90‑day figure as a rule of thumb, not our measurement. Our own index is a rolling observation window, so the longest continuous runs we have on record currently top out around 28 days. Even at that horizon the pattern is loud. Ads like SmartAsset's "How Can I Avoid Paying Taxes on IRA Withdrawals?" on Outbrain and a wall of "Try next‑gen hearing aids" placements on the Microsoft Audience Network have held their slots for the full 28‑day window we can see.
4. Where it sits in the supply chain. The path from impression to advertiser passes through SSPs, exchanges, and trackers, and reading that chain tells you a lot about budget and sophistication. See The Native Ad Supply Chain, Explained (With Real Traces).

That layer of analysis on top of the raw archive is what separates a library from an ad intelligence platform. The full discipline (what data exists, how to read it, and what it is worth) is covered in What Is Ad Intelligence? The 2026 Guide to Competitive Ad Data. If your specific question is "who is buying ads on this site," that has its own playbook in Who Is Buying Ads on a Website?.

How to actually use ad transparency#

A workflow that holds up regardless of vertical:

• Start with the platform library that matches the channel. Researching a brand's social? Meta's library first. Search and YouTube? Google's. Native and content recommendation? You need an independent native library, because no official one exists.
• Search the advertiser, then search the angle. Pull every active creative for a competitor, then search the offer or hook across advertisers to see who else is running the same play. The second search is where you find rising competitors you did not know existed. The IRS‑forgiveness angle above, for instance, is not one brand; it is a whole template that copycats clone the moment it works.
• Filter for longevity. Sort by run time and ignore the brand‑new creatives. The ads that have survived weeks or months carry the budget, and they are the only ones worth modeling.
• Follow the click before you trust the brand. On any native or programmatic placement, trace the destination. The landing page tells you the real advertiser, the offer, and often the funnel, none of which the ad card reliably shows.
• Save and watch, do not just browse. Transparency data is most valuable over time. A creative that appears, scales, and then survives is a signal. A single snapshot is noise.

This is also where the gap between "browse the official archive" and "operate on the data" gets obvious. Browsing is free and shallow. Operating means classifying the supply chain, resolving the real advertiser, tracking longevity, and pulling it all through an API or MCP into your own tooling, then turning what you learn into action with features like Creative Studio, Optimize, and Copy DNA.

Where OpenAdLibrary fits#

OpenAdLibrary exists for the half of the advertising world the official libraries skip: live public native ads on the open web. It is an open, low‑cost alternative to the $80 to $400 per month incumbents, with a free tier that lets you browse 200 ads with no card. It captures the real creative at full quality, classifies the ad‑tech supply chain, traces each click to the landing page so you can see the real advertiser, and surfaces the longevity and spread signals that reveal which ads are actually winning. The numbers behind it (589,036 creatives, 25,933 advertisers, 42 networks, 5.4 million observations as of June 2026) are the record the native ecosystem never had.

If you have hit the wall where Google's and Meta's libraries stop, which is the entire native ecosystem, this is the missing piece. Start free and browse 200 live native ads, no card required.

Keep going#

This page is the center of a cluster. To go deeper on any one thread:

• What Is a Native Ad Library and Why One Didn't Exist Until Now explains the structural reason the native ecosystem had no archive, and how that changed.
• What Is an Ad Transparency Tool? Open Ad Libraries Explained compares the official libraries field by field, including what each one omits.
• What Is Ad Intelligence? The 2026 Guide to Competitive Ad Data is the practitioner's guide to turning transparency data into decisions.
• The Native Ad Supply Chain, Explained (With Real Traces) maps the path from impression to advertiser with real examples.
• How to Identify the Ad Network Behind Any Ad is the hands‑on method for attributing any placement to its network.