Still feel like we're talking past each other a bit here. Some portion of your UI logic is going to have to translate that data into the corresponding UI output, whether it be HTML elements or Android Views or iOS NSWhateverThingsTheyUse. With a virtual DOM, that's a two step process: your component is responsible for doing the "data -> desired UI structure" translation, and then the VDOM layer is responsible for translating that into the actual UI pieces. If you've got {likes : 42}, React has no idea what a "like" is, what it means to have 42 of them, or what it should do with that information. Your code has to tell it what that means in terms of something to actually draw.