They mean that it's fraud to mislead people by gaming their assumptions about performance indicators.

Specifically, they're saying that paying for one kind of popularity (e.g., sales of widgets or good reviews on Amazon.com) is fraud when that popularity is meant to trick people into believing that it implies a subtly different kind of popularity (e.g., sales of widgets from independent third parties or good reviews on Amazon.com from unbiased reviewers).

Sure, in a certain sense. But the subject of the article is something fundamentally different in lots of the nuanced ways that a useful, complete mental model of the world so annoyingly and doggedly hinges on.

This isn't advertising. The article talks about a specific action, which is paying someone to buy cheap things from you, so you can tell investors that you have a lot of sales.