Ad Fraud And Synthetic Traffic
Synthetic traffic is the least glamorous consequence of cheap generation, and probably the most expensive. Manufacturing a site, filling it with plausible articles and selling the impressions is now a weekend of work.
The economics that always drove ad fraud have not changed. What has changed is that the cost of the deception fell faster than the cost of detecting it.
How Synthetic Traffic Reaches A Media Plan
Through the long tail. A campaign buying broad reach programmatically ends up on thousands of domains nobody named, and the mechanics that place it there reward volume over provenance.
The generated site is designed to pass the checks that exist: it has a privacy policy, plausible content, a category, and a history of serving ads without complaint.
Why Verification Struggles
Because most verification was built to answer whether a human saw the ad in a viewable position, not whether the page had any reason to exist. A site can be entirely synthetic and still serve viewable impressions to real browsers.
The response that works is older than any of this: buy from a shorter list. An inclusion list of known publishers costs more per impression and removes an entire category of loss, which is a trade most Indian advertisers can make and many have not.
The cheapest thousand impressions in the market are cheap for a reason, and the reason is not efficiency.
What Advertisers Should Ask This Quarter
For the actual domain list, in full, not the top twenty. For the share of spend on sites first seen in the last six months. And for the agency's position on inclusion lists, which is a philosophy question disguised as a settings question.
None of this is new advice. It is simply more urgent now that the supply of convincing fake inventory is effectively unlimited.
Related: whether AI media buying is efficient or opaque.
More in The AI Shift.