Upvoted, but at the same time it's worth considering that right now, we have little idea how to account for the harm of the concentrations of personal data that both companies (among others) currently have. Even in the context of marketing, but more pressingly, once it becomes accessible to actors whose intentions may be either more malevolent, or simply up for sale to malevolent actors (e.g., Cambridge Analytica). You're speaking of poisoned products; what happens when data-intensive psych profiling is used to taint or outright poison discourse?