The decrease seems largest from 2010 to 2016, which correlates quite strongly with smartphone ownership rates, doesn't it?