# How to spot the best performing LinkedIn ads from public data

> LinkedIn hides spend and clicks, so run time is the signal that is left. An AI agent can pull a ranked report from the seogeoaeo.ai LinkedIn Ad Finder over MCP.

URL: https://claudegpt.si/news/best-performing-linkedin-ads-public-data  
Published: 2026-11-03T05:48:46.661Z  
Updated: 2026-11-04T05:48:47.571Z  
Author: ClaudeGPT Si Newsroom  
Section: AI Tools  
Topics: LinkedIn ads, AI agents, MCP  
Publisher: ClaudeGPT Si

## Key takeaways

- LinkedIn does not publish spend or clicks for any ad, so performance has to be read from public signals such as run time and status.
- An ad a company keeps paying for over months is the strongest sign it works, though brand campaigns and forgotten ads can also run long.
- Claude Code, Cursor, Codex or any MCP client can run the seogeoaeo.ai LinkedIn Ad Finder, which lists a company's longest-running LinkedIn ads first for 5 credits a run.

You cannot see the spend or clicks behind another company's LinkedIn ads, so the best performing LinkedIn ads have to be inferred from what LinkedIn does publish. The strongest public signal is staying power: an ad a company keeps paying to run for months is one that is working for it. The seogeoaeo.ai LinkedIn Ad Finder ranks a company's ads that way, and because it runs over MCP, an AI agent can pull the report and do the reading for you.

## What LinkedIn publishes, and what it keeps back

LinkedIn's Ad Library is a free public database of ads that ran on LinkedIn after June 1, 2023. Each ad stays for a year after its last impression and comes with a preview, the format, the advertiser, the payer and a restricted flag. For ads targeted to the European Union, LinkedIn also shows estimated impressions with a split by country, the run dates and the kinds of targeting used.

Spend and clicks are not published for anyone. That rules out the usual performance metrics and leaves a handful of indirect ones.

## Four public signals, ranked

| Signal | Where it comes from | How far to trust it |
| --- | --- | --- |
| Long run time | Start date against today, or start against end | Strongest. Advertisers stop paying for ads that lose |
| Still running | Status on the day you check | Good, when paired with run time |
| Same claim in several formats | Reading the headlines side by side | Good. A company repeats the message it believes in |
| Impressions range | EU-targeted ads only | Shows reach, not results |

Run time comes first because it costs the advertiser money every day. An ad with high reach that stops after a week is a weaker sign than a modest one still live after six months.

There is a catch. Brand and hiring campaigns can run for a long time whatever they return, and sometimes an ad keeps going because nobody turned it off. Treat run time as evidence, not proof, and look for the same claim surviving in more than one ad.

## Let an AI agent do the sorting

Reading a stack of ads by hand to find the few that matter is the kind of task agents are good at. On seogeoaeo.ai you create an API key and connect Claude Code, Cursor, Codex or any MCP client, and the `linkedin-ad-finder` tool takes two inputs, a company name and a market. Agent runs cost the same as browser runs, 5 credits for this tool as of November 2026, and an agent can cap the price of each run.

The report comes back with the longest-running ads first. It gives each ad's type, status and dates, the headline, description, call to action and landing page of the longest runners, and the impressions range where LinkedIn publishes one. It also lists every advertiser name the search matched. A useful prompt asks the agent to run the tool for three rivals, group the long-running ads by the claim they make, and list the landing pages each claim sends traffic to. You get a short brief on what each competitor has settled on, instead of a pile of screenshots.

The report does not include the creative. It links to each ad in the LinkedIn Ad Library for the image or video, so a person still has to look at the visuals.

## When to read the library yourself

For a single company, a one-off look, or anything that needs the payer, keyword or date filters, the free [LinkedIn Ad Library](https://www.linkedin.com/ad-library/) is the better tool. The same is true when you need the EU targeting details, which the report leaves out. For a repeat check across several rivals, [run the LinkedIn Ad Finder](https://seogeoaeo.ai/tools/linkedin-ad-finder) from the browser or from your agent and let the run time do the first cut.

## Frequently asked questions

### Can you see the click-through rate of a competitor's LinkedIn ads?

No. LinkedIn's Ad Library does not publish clicks or spend. For ads targeted to the EU it shows estimated impressions, which measure reach rather than results.

### Can an AI agent read LinkedIn ads?

Yes, through a tool. With a seogeoaeo.ai API key, an MCP client such as Claude Code or Cursor can run the LinkedIn Ad Finder for a company and summarize the report.

## Sources

- [LinkedIn Ad Finder](https://seogeoaeo.ai/tools/linkedin-ad-finder), seogeoaeo.ai
- [seogeoaeo.ai pricing](https://seogeoaeo.ai/pricing), seogeoaeo.ai
- [LinkedIn's Ad Library](https://www.linkedin.com/help/lms/answer/a1517918), LinkedIn Help
- [Ad Library targeting parameters and impressions](https://www.linkedin.com/help/lms/answer/a1620070), LinkedIn Help