Know which clicks are actually human.
Skyfense identifies the paid campaign behind a visit, evaluates that session for automation and fraud, and records whether the click was valid. This connects ad performance to the same evidence used to protect checkout.
Click IDs persist in sessionStorage for the life of the tab, so a checkout page still knows the campaign that paid for the visit.
Attribute it.
Assess it. Act on it.
Identify the source
On landing, Skyfense reads click IDs such as gclid, fbclid, msclkid, and ttclid, plus campaign UTM values. If an ID is stripped, the referrer can identify the network.
Assess the session
The attribution follows the tab through the funnel. Bot, automation, ML, behavior, IP, and checkout signals are checked together rather than with a separate ad-fraud score.
Report the result
Each attributed session receives a valid or invalid assessment with its network, campaign, click ID, and reasons. The dashboard aggregates those results by network.
Don't let a bot
teach your ads.
When enabled for a workspace, the guard loads before ad-network snippets and watches their event queues: Google dataLayer, Meta fbq, Microsoft uetq, and TikTok ttq.
If the session is blocked, calls that would report a click or conversion are silently suppressed. The page still works, while fraudulent behavior stays out of the optimization data used by the ad platform.
This safeguard is opt-in because a false positive can affect legitimate reporting. Merchants should validate detection before enabling it.
No second model.
The same evidence.
Known crawlers, fake Googlebot traffic, scrapers, scanners, and datacenter browsers are classified before the ad assessment.
Selenium, Puppeteer, Playwright, headless artifacts, and the behavioral model feed into click validity.
Review network, campaign, click ID, valid/invalid status, reasons, and the linked session forensics.
Ad spend protection questions
What marketers ask about catching invalid clicks and protecting ad budget.
How does Skyfense know a click came from an ad?
It reads the ad tracking code on the landing link, such as Google gclid, Meta fbclid, Microsoft msclkid, or TikTok ttclid. If that code is missing, it can recognize the ad network from the referring website.
Will this affect my real ad reporting?
The pixel guard only holds back events for sessions judged fraudulent, and it stays off until you switch it on. That way you can confirm the detection fits your traffic before anything is suppressed.
Which ad platforms are covered?
Google Ads, Meta, Microsoft/Bing, and TikTok today, plus generic UTM campaign links from any source.