This seems broadly to pass the Kant test on morality, so it could be said to be well-grounded.
Something is immoral if it is contradictory when generalised -- that was the gist of Kant's view. And real manipulation and persuasion seem to fail: if everyone is allowed to manipulate everyone else, everyone loses their own control of themselves -- in which case, how can they manipulate someone else? There is a contradiction.
The two basic checks the article proposes -- "Will I use the product myself?" and, "Will the product help users materially improve their lives?" -- are sort-of rough ways of testing for such Kantian contradction.