The only example of actual increase of use in a particular feature is this one: https://miro.medium.com/max/2400/0*24guSnnmXQu2zi3n https://miro.medium.com/max/

The only example of actual increase of use in a particular feature is this one: https://miro.medium.com/max/2400/0*24guSnnmXQu2zi3n https://miro.medium.com/max/2400/0*24guSnnmXQu2zi3n (the Google Translate example) And in that it's not even just about icon vs icon + text, but two very different kind of icons... One is a snake (?) and the other one a pen. Edit: Also, want to answer directly to "the increase in use of features ": that also doesn't necessary means it's good for the user. Maybe they don't actually have any use of it, and what they could see before on the main page, they now try to find. Point being, designing UX after metrics is a bit pointless, as it doesn't show anything that you should really care about when solving someones problem.