Hacker News Reader

I looked it up and it's not annualized, it's year over year. Therefore very sensitive to chosen start and endpoints.

OK, but with year over year, it makes exactly zero sense to say that the May-to-May change of 17% was "on top of" the April-to-April change of 15%.

Maybe they meant it’s “on top of” it in a table of YoY changes by month sorted in descending order ¯\_(ツ)_/¯