LinkedIn Ads Library: The Complete Guide to Competitor B2B Ads

Most guides to LinkedIn's Ad Library describe it as a thin tool: public, searchable, shows you the creative, does not show you spend. That is accurate for a US ad and badly wrong for a European one, and the difference is the single most useful thing to know about it.

We spent 2026-07-30 in the library, anonymously, comparing what the same advertiser's ads expose in the United States against the same advertiser's ads in Germany. The US ad gave us four fields. The German ad gave us a bracketed impression total, a per-country impression split, an exact date range, and an itemised targeting disclosure. Same library, same advertiser, same day.

This guide covers what the library exposes, the regional asymmetry and how to use it deliberately, the duplication trap that makes result counts meaningless, and a workflow that ends in tests rather than screenshots.

What it is, and what it costs you

linkedin.com/ad-library is fully public. No LinkedIn account, no advertiser access, no rate limit we hit. LinkedIn describes it as offering "transparency in advertising by providing a searchable collection of ads."

The search bar exposes six filters: Company or advertiser, Payer, Keyword, Country, Date, and a More menu. Five of those are what you would expect. The Payer filter is the one almost nobody uses, and we will come back to it, because it is the most powerful field in the tool.

Per LinkedIn's own documentation, the baseline data shown for any ad is: ad preview, ad format, advertiser name, payer name, and whether the ad is restricted. Ads appear "shortly after the ad receives its first impression", and — critically — remain in the library for one year after their last impression.

The retention rule, and why it is a gift

Read that retention rule carefully, because it is one year from last impression, not one year from when the campaign was paused or created.

Think about what that does to the population of ads you can see. A two-week experiment that ran in early 2025 and stopped is gone. An ad that has been running continuously since 2024 is still there, and its retention window keeps rolling forward every day it serves. The library does not sample LinkedIn advertising evenly — it systematically preserves long-running ads and deletes short-lived tests.

For a compliance archive that is a distortion. For a B2B researcher it is close to a gift, because in paid media longevity is the most reliable public proxy for performance that exists. Nobody keeps paying to serve an ad that loses, and B2B budgets are scrutinised harder than most. An ad that has been live for fourteen months has survived more audience saturation than any test you will run this quarter.

So the highest-value question in the library is not "show me everything this competitor runs". It is "which of this competitor's ads have been running the longest" — and for EU-targeted ads, the date range needed to answer that is printed on the page.

The regional asymmetry — the actual exploit

Here is the comparison that matters. Two ads from the same advertiser, opened on the same day.

A US-targeted ad detail page showed exactly this:

About the ad
  Video Ad
  Advertiser      Google
  Paid for by     Essencemediacom LLC

Four fields. No dates, no impressions, no targeting.

A Germany-targeted ad from the same advertiser showed this:

About the ad
  Single Image Ad
  Advertiser      Google
  Paid for by     ESSENCE GLOBAL LIMITED
  Ran from        Jul 16, 2026 to Jul 29, 2026

Ad Impressions
  Total Impressions   300k-500k
  Germany             100%
  India               < 1%
  United States       < 1%
  Switzerland         < 1%

Ad Targeting
  Language    Targeting includes English
  Location    Targeting includes Germany
  Targeting parameter | Targeted | Excluded
    Audience
    Demographic
    Company
    Education
    Job
    Member Interests and Traits

That is a different product. The reason is regulatory: the EU's Digital Services Act requires ad repositories to disclose impressions, targeting and run dates for ads served in the EU, and LinkedIn's documentation confirms that for EU-targeted ads the library displays "information about ad impressions, ad targeting, and the dates the ad ran." Ads outside that scope get the minimum.

So the workflow implication is direct: if your competitor advertises in any EU country, research them there — not in your home market — and you get impressions, dates and targeting for free. For B2B software, where almost every serious advertiser runs at least a DACH, France, Benelux or Nordics campaign, this is available far more often than people assume. You are not looking at their US campaign, so read it as directional rather than identical. But a competitor's positioning, offer structure, targeting philosophy and creative format rarely differ wildly across markets, and the EU version is documented.

Four things that specific disclosure gives you that no third-party scraping tool sells:

1. Impression scale, bracketed. 300k–500k is a wide band, so it will not rank two ads precisely. What it does do is separate a real flight from a token one, and it lets you sanity-check whether a competitor is committing budget to a message or merely testing it.
2. Delivery spill. Germany 100%, India <1%, US <1%, Switzerland <1%. Those sub-1% entries are leakage outside the intended market. Consistent spill across a competitor's ads suggests loose location settings — expanded rather than exact location targeting — which is a real operational tell.
3. The language-versus-location split. This ad targets Germany but its language targeting includes English. That is a deliberate choice to reach English-speaking professionals in Germany rather than German-language ones, and it tells you something specific about who the competitor thinks the buyer is. It is exactly the kind of decision that never shows up in the creative.
4. Exact run dates. Which makes the longevity proxy computable rather than guessed.

The duplication trap

Searching one large advertiser in the US returned 1,124 ads match your search criteria. That number is close to meaningless, and reading it as creative volume is the most common mistake made with this tool.

Scrolling those results, the same handful of creatives repeated over and over — the same headline, the same opening copy, sometimes five or six times in a single screen. LinkedIn explains why on the detail page itself: "Please note that this ad may have multiple versions. Some versions may appear different on mobile or desktop screens with different sizes."

Each version is its own library entry with its own ad ID. So 1,124 entries might represent thirty or forty distinct creative concepts. Before you conclude anything about a competitor's creative velocity, deduplicate by headline plus first line of body copy. Otherwise you will report that a competitor is running a thousand ads when they are running forty ads in twenty-five formats, which leads to exactly the wrong strategic conclusion.

Two things visible in the library that most people miss

Unrendered personalisation macros. Several results displayed literally as %FIRSTNAME%, work with us at Google and %FIRSTNAME%, explore jobs at Google that match your skills. The library renders the raw macro rather than resolving it. That is genuinely useful intelligence: it confirms the competitor is running dynamic ads with profile-based personalisation rather than static sponsored content, and it shows you the exact template. You are reading their creative source, not their output.

Thought Leader Ads are labelled. Some entries appear under a person's name and headline — "Prof. Dr. Yasmin Weiß, Professor of AI at work..." — with "Promoted by Google" beneath. These are Thought Leader Ads: a company paying to amplify an individual's post. They are worth separating in any analysis, because they are a different play with different economics from brand-account sponsored content, and in B2B they are increasingly where the budget goes. The EU version of one of these carried the full impressions and targeting disclosure, so you can measure the play.

The Payer filter, used properly

Payer is the legal entity that paid, and it is searchable in its own right. In our two examples the same advertiser's US ad was paid for by Essencemediacom LLC and its German ad by ESSENCE GLOBAL LIMITED — the same agency network, different legal entities per region.

Three ways to use that.

• Find an agency's whole book. Search a payer entity and you surface every brand that agency buys for on LinkedIn. If you are pitching against an agency, or trying to understand which shop is winning in your category, this is a client list you cannot get anywhere else.
• Detect an agency change. If a competitor's payer entity changes between older and newer ads, their agency changed. That usually precedes a strategy change, and it is a leading indicator you can watch for free.
• Map how a global account is split. Different payer entities per market tell you which agency office owns which region — and therefore where creative consistency is likely to break down, which is where a focused competitor is easiest to out-execute.

A workflow that produces tests

Step 1 — Build the advertiser list before you open the tool. Five to fifteen names: direct competitors, the category leader, and two or three adjacent companies selling to the same buyer for a different reason. Starting from a keyword search is how you end up with four hundred irrelevant ads.

Step 2 — For each name, run the search twice: your market, then an EU market. The home-market pass tells you what they say to your buyer. The EU pass tells you impressions, dates and targeting. Germany, France and the Netherlands are the usual hits for B2B software.

Step 3 — Deduplicate before counting anything. Headline plus first line of copy. You are building a list of concepts, not entries.

Step 4 — Sort the EU results by run length. First-shown to last-shown. The longest runners are the survivors and deserve most of your attention; everything else is inventory.

Step 5 — Decompose survivors on five axes. Hook in the first two lines, format (single image, video, document, carousel, Thought Leader Ad), the specific claim or offer, the call to action, and — from the EU targeting block — who it was aimed at. Five columns in a sheet. Patterns emerge at around fifteen rows.

Step 6 — Read the targeting against the creative. This is the step people skip and it is where LinkedIn's library beats every scraping tool. When you can see a survivor was aimed at English-speaking professionals in Germany with specific job and company parameters, its creative choices stop looking arbitrary. Ads that look strange usually look strange because you are not the target.

Step 7 — Convert patterns into one hypothesis per test. "Four of six survivors lead with a named customer problem rather than a product capability" is a hypothesis with a metric attached. Write it before you brief anything, or the research quietly becomes justification for whatever you were already going to make.

What the library cannot do

Worth stating plainly so you stop looking for it.

No spend, ever — impressions for EU ads are the closest proxy and they are bracketed. No CTR, conversion rate, or lead volume. No indication of which entries were A/B variants of each other, which is precisely what you would need to learn what a competitor learned. No landing pages beyond what is visible in the creative. No official API, so extraction is manual or scraped, and scraping a login-free page still carries terms-of-service risk worth checking. No US targeting or impressions data, and no sign that will change without US regulation.

And the biggest one, which is structural rather than a missing feature: the library tells you what a competitor did, never what it earned. Longevity is a good proxy and it is still a proxy. A well-funded competitor can run a mediocre ad for a year.

Where this fits

Competitor research fails in a predictable way. Someone spends a good afternoon in the Ad Library, produces a document full of screenshots and observations, shares it, and nothing in the ad account changes — because the research lives in a slide deck and the performance data lives in Campaign Manager, and nothing joins them.

The join is the work. A pattern is only worth the afternoon if it becomes a funded test with a named metric, and if the result of that test changes what you research next. Everything in the workflow above is designed to end at a hypothesis for exactly that reason.

If you are running the same play on other platforms, our guide to the TikTok ads library covers the equivalent transparency archive there — including the same one-year-from-last-impression retention rule, which turns out to be the shared foundation of competitor research across both platforms. And our audit of TikTok Creative Center covers the curated, opt-in counterpart, which answers a genuinely different question.

All library behaviour, field structures, filter options and disclosure differences in this article were verified by browsing linkedin.com/ad-library anonymously on 2026-07-30, comparing US-targeted and Germany-targeted ads from the same advertiser. Retention and appearance timings are from LinkedIn's own published help documentation.