Google Ads cloaking: AdsBot, the Circumventing Systems policy and how a cloaker is configured

Google Ads cloaking is the practice of showing Google's review systems a different landing page than the one buyers see. Google is more explicit about it than any other ad network: cloaking is listed by name in the Circumventing Systems policy, and a confirmed violation can suspend the account without a prior warning. This guide is not about "beating" that policy. It explains what Google's crawlers actually do on a landing page, which signals a filter can use, how a KLO campaign is configured for Google traffic and how to verify the result with Google's own tools.

The Facebook version of this guide is here; the two networks check landing pages in very different ways.

What Google calls cloaking and what it costs

The Circumventing systems policy has a dedicated Cloaking section. Google's wording: cloaking is showing different content on your website to different people or to Google in order to hide things that might break Google Ads' rules. Showing slightly different content is fine, for example different languages, offers by region or a lighter page for slow connections, but the promoted product or service has to be the same for everyone. Click trackers and redirects are allowed as long as they are not used to hide a policy violation.

The examples Google gives are the ones a cloaker produces by design: showing Google a landing page that complies with the policies while showing people different content, an intrusive interstitial that blocks Google's access to the page, leading users to a website on an entirely different topic while hiding this from the crawlers, and sending users to a policy-violating page even within the same domain.

The important word in the policy is egregious. Circumventing systems is one of the policies Google treats as an egregious violation: in Google's own words, accounts are suspended upon detection and without prior warning, the advertiser is not allowed to advertise with Google Ads again, and accounts are only reinstated in compelling circumstances. In practice the suspension extends to related accounts as well.

Two practical consequences. First, Google does not need to prove that the offer page violates another policy: the mismatch itself is the violation. Second, the review is continuous. Landing pages are re-crawled during the life of a campaign, not only at approval.

Google's crawlers on a landing page

Google documents which crawlers check ads and how they behave. For a landing page filter the relevant ones are the special-case crawlers.

Google also publishes the IP ranges of its crawlers as JSON files. As of September 21, 2026 the files contained 317 prefixes for common crawlers, 272 for special-case crawlers including AdsBot, and 496 for user-triggered fetchers. The recommended verification is reverse DNS: the name must end in googlebot.com, google.com or googleusercontent.com, and the forward lookup must return the same IP.

What this means for a filter: a declared AdsBot from a published prefix is the easy part. A check that arrives with a normal Chrome user agent from a Google Cloud address is the hard part, and it is why network data has to cover Google's cloud ranges, not only the crawler ranges.

Click IDs: gclid, gbraid and wbraid

With auto-tagging on, every click from Google Ads carries a click ID. On most platforms it is gclid. On iOS 14.5 and later, where app tracking restrictions apply, Google uses gbraid for app-to-web and wbraid for web-to-app measurement instead. A filter that requires exactly gclid sends part of the real iOS traffic to white.

In KLO the click ID rule for Google accepts all three by default. The strict variant, which demands gclid only, exists for campaigns that never run on iOS.

Setting up KLO for Google Ads

Step 1. White page on the ad's topic

Google scores the relevance between the ad and the destination. The white page must deliver what the ad promises, have contact details and a privacy policy, and load fast on mobile: AdsBot-Google-Mobile checks the mobile version separately. The white page guide has the checklist.

Step 2. Campaign rules in the cabinet

Source Google Ads, countries and devices from the campaign settings, the click ID requirement with gclid, gbraid and wbraid, the white and offer addresses. Warm-up is mandatory for a new campaign: the first AdsBot checks arrive right after launch and after every ad edit.

Step 3. Connection

PHP loader, WordPress plugin, Keitaro module, SDK or Cloudflare Worker. For Google traffic the Worker option is particularly convenient: the decision is made at the edge before your hosting, and there are no cloaking files on the server that a manual review could find.

Step 4. Check what Google sees

Search Console has a URL Inspection tool. The Test live URL button fetches the page with a Google crawler and shows the HTML and a screenshot it received. If the filter is configured correctly, you see the white page. This is the official way to confirm that Google's crawler gets what you intended for it.

Step 5. Monitoring

After launch watch the visits from Google networks in the KLO statistics: their rhythm shows when a review has passed. A sharp rise in visits without a click ID from one subnet usually means a re-review or a spy tool.

Mistakes specific to Google

• Requiring gclid only. iOS traffic with gbraid and wbraid goes to white and the campaign looks like it does not convert on iPhone.
• Blocking AdsBot in robots.txt. AdsBot ignores the global rule, and a specific disallow for AdsBot-Google is itself a review flag. Let the filter answer with white instead.
• A slow white page on mobile. AdsBot-Google-Mobile is a separate check with its own quality score. Test the white on a phone.
• Different tracking on white and offer. The white page without the conversion tag that the campaign uses is a mismatch on its own.
• Trusting the crawler IP list alone. Manual and semi-manual checks come from Google Cloud and from ordinary networks. Combine network data with the click ID and device rules.