Point 3 seems to be a case where correlation does not imply causation. Basing an advertising campaign with an oblique reference to the issues at Uber* doesn't mean the campaign is effective.

* most people I talk with are only vaguely aware there are issues at Uber, including many drivers, Uber users, and non-Uber users.