This I think is one of the most important questions of personalization and recommendation systems. There’s nothing better than giving someone a true and thoughtful reason for why you made a recommendation. And it always comes down to context. When I tell someone about a book, for example, I don’t just show them the title. I explain why I thought they may particularly like it, and I often nuance it “if you don’t like X, Y, Z, you may not like this one, but I still think you’ll enjoy A, B, C”. When I get that enough from someone, then over time they can just say “check out X” and I’ll assume the rest. This is why brands like New York Times still count for me because I authentically trust what they stand for. But this is very hard to do with algorithms so they never start that way and start earning trust. The ones that do will win in the long run I think!