I think it's this: > The study’s authors—[…]—found that users were 45% more likely to click on results that were ranked purely by relevance, rather than as Goo

I think it's this: > The study’s authors—[…]—found that users were 45% more likely to click on results that were ranked purely by relevance, rather than as Google ranks them now, with its own services displayed prominently. If users are clicking on results 45% less often, that suggests they aren't getting results that are as useful to them. From that it seems fairly reasonable to claim that this is harmful to consumers.