OpenAdLibraryOpenAdLibrary
Ad Transparency & Supply Chain

How Scam Ads Slip Past Review: Cloaking, Swaps & Weak Spots

Ad review has gaps, and some scam campaigns exploit them. Possible failure modes such as cloaking, post-approval swaps and redirect switches, and the signs that can help expose them after approval.

Editorial illustration: How Scam Ads Slip Past Review: Cloaking, Swaps & Weak Spots

An approved ad can still be deceptive. Possible explanations include a review error, content changed after review, or different content being shown to reviewers and consumers. Platforms' review and ongoing enforcement processes vary. Approval should not be treated as independent verification of an advertiser's claims.

This article walks through each technique, explains why native ad networks get hit especially hard, and shows how to recognize — and document — a scam ad that made it through review.

Ad review is a snapshot. A scam campaign is a movie#

Platforms may combine automated checks, human review and enforcement after publication. The evidence here does not establish the exact checks, triggers or monitoring frequency used by every major network.

Advertiser-controlled destinations and redirects can change after an initial review. Such changes can create a gap between an earlier decision and a later consumer experience, but they may also trigger platform monitoring or enforcement. Preserve the time and context of each observation.

Technique 1: cloaking — the reviewer never sees the scam#

Ad cloaking means serving different content to reviewers than to real users. The mechanics are unglamorous: the landing page, or a router sitting in front of it, checks each visitor against lists of known data-center and ad-network IP ranges, then checks geo, device type, user agent, and sometimes behavioral signals. Visitors who look like reviewers get a "white page" — a bland, compliant article or storefront. Visitors who look like the target audience get the money page: the fake celebrity endorsement, the miracle cure, the investment scheme.

Cloaking is sold as a commercial service with monthly pricing, dashboards, and support channels, which tells you how mature the market is. Countermeasures exist — networks re-crawl from residential-looking IPs and randomize their review fingerprints — but it is an arms race in which the defender has to win every fetch and the cloaker only has to win at review time. We cover the detection side in depth in how ad cloaking works and how evidence exposes it.

Technique 2: the post-approval swap#

A post-review content change can replace a previously reviewed page with materially different claims. A dated capture of each version helps document that change. Do not assume a platform never rechecks the destination.

A variant targets the ad side. Some networks allow edits to headlines, thumbnails, or destination URLs on live campaigns with lighter-touch re-review than a fresh submission. Operators probe exactly how much they can change before a full re-review triggers, then stay under that line. The ad that got approved and the ad that runs at scale are technically the same campaign object — and materially different artifacts.

Technique 3: redirect chains that move after approval#

Almost no performance ad points directly at its final page. The click travels through a redirect chain: the network's click handler, then one or more advertiser-controlled tracking domains, then the lander. The network reviews the destination it can see at submission — but the middle of the chain belongs to the advertiser, and a tracker is just a server answering "where should this click go?" That answer can change at any moment, or vary per visitor.

The per-visitor part is the potent bit. A tracker can split traffic by geo, device, carrier, or time of day: a reviewer or monitoring service fetching from a US data center gets the compliant path, while mobile users in the target geo get routed to the scam. The approval was honest, in a narrow sense. Everything after it was theater.

Technique 4: burner accounts and the economics of volume#

Some abusive advertisers may return through replacement or compromised accounts. The evidence here does not quantify account costs, detection time or the profitability of doing so.

This is why "why doesn't the network just ban them?" misses the shape of the problem. Networks ban them constantly. The countermeasure that actually bites is fingerprinting repeat offenders across accounts — payment instruments, creative assets, landing infrastructure — and networks have invested heavily there. But the contest is adversarial, and the attacker iterates faster than the defender's release cycle.

Why destination context matters in native ads#

Advertorials and pre-landers can be used legitimately or deceptively. Their format alone does not establish fraud. Assess the truth of the claims, source identity, disclosures and destination behavior.

Demand resale adds another layer. On parts of the native supply chain, an ad can enter through a demand partner and surface on a network that never had a direct relationship with the advertiser, diluting accountability at each hop. And the vertical mix concentrates temptation: across OpenAdLibrary's index of 725,000+ live native ad creatives (June 2026), health and finance are the two largest verticals at roughly 24,000 classified creatives each — exactly the categories where miracle cures and get-rich schemes prefer to hide. We break down the wider taxonomy in Ad Fraud in Native Advertising.

What review actually catches#

None of this means review is useless. Modern review stacks catch an enormous amount: image classifiers flag celebrity faces and before/after shots, text models flag miracle claims and banned categories, landing scanners follow declared destinations, verification programs demand business documents in sensitive verticals, and payment-risk systems kill accounts before the first impression serves. The obvious, lazy scams die at submission — which is why you mostly see the survivors and wonder how they got through.

The examples explain possible failure modes, not their prevalence. We do not have a representative dataset of submissions, approvals and violations from which to calculate review accuracy.

How to spot an approved scam ad in the wild#

A few signals separate scam ads from merely aggressive ones:

Investigate an unverified endorsement, an unexpected destination, a recently registered domain or inconsistent advertiser identities. Each is a lead, not proof of a scam: agencies, affiliate relationships and legitimate tracking can also produce differences. Corroborate the claim through authoritative sources and dated destination evidence.

Verifying most of these requires seeing the click path and the landing page, which is exactly what an independent ad intelligence index is for. OpenAdLibrary records each native ad with its resolved advertiser, redirect trail, and captured landing page — over 1.3 million landing captures as of June 2026 — so you can check what an ad actually resolved to without clicking it yourself. If the ad borrows a brand's identity outright, that is its own playbook: see Copycat Landing Pages and Trademark Infringement in Ads.

Report it with evidence that sticks#

Include the creative, publisher URL, timestamp and time zone, observed destination and a clear explanation of the suspected violation (see our guide to reporting a scam ad). These details can support an investigation but do not guarantee a response or removal. US consumers can also submit reports at ReportFraud.ftc.gov.

Approval is not an endorsement. It is a snapshot that a motivated adversary controls. Once you internalize that, scam ads stop being mysterious — and their tells become much easier to read.

Frequently asked questions

How do scam ads get approved by ad networks?
Possible causes include review errors, destinations changed after review and content that differs between reviewers and consumers. Platform processes vary, and approval is not verification that an ad's claims are true.
What is ad cloaking?
Cloaking is serving different content to ad reviewers than to real users. A router checks each visitor's IP, geo, device, and user agent; suspected reviewers get a bland compliant page while target users get the real landing page — often a fake endorsement or miracle-cure funnel. It is sold openly as a commercial service and banned by every major network.
Why do scam ads keep coming back after being banned?
Replacement or compromised accounts can allow abusive activity to recur. The evidence here does not establish typical account costs, detection times or profits. Report repeat patterns with dated evidence rather than assuming every case uses the same method.
Are native ad networks worse at catching scam ads?
This article does not establish a comparative review-accuracy ranking. Native formats and resold demand can create context to inspect, but neither proves that a network is worse at detecting scams.
How can I check whether an ad is a scam without clicking it?
Inspect available dated captures and verify claims or endorsements through authoritative sources. Unexpected domains or recent registrations are investigation leads, not proof. Avoid interacting with a suspected scam or clicking its live ad to gather evidence.
The OpenAdLibrary Team
Written byThe OpenAdLibrary Team
Ad intelligence & native advertising research

We build OpenAdLibrary, the open ad-transparency platform. Every day our systems capture live native ads across Taboola, Outbrain, MGID, Revcontent, Teads, Yahoo and MSN, identify the real advertiser behind each one, and follow the click to its landing page. These guides distill what we see in that data so you can research the market faster.