In reality it is difficult to predict what will gain traction. I don't think this is even related to whether a product is useful or not. If I build a product that there are a million of already, it may be useful sure, but will it be picked up by the masses?

At a company I'm at now we just implemented a heart button on pages without it actually doing anything, except we track if people click on it or not. This is a smaller version of the same validation. If it's justified, we will build out the feature of liking things. Building out entire castles and worlds that no one lives in is very tempting for a passionate developer.