This seems like a pretty shaky data foundation for an article - the "happiness" rating oscillates between 7 and 9 on a 10 point scale, and they don't explain why they picked the timeframes they did (which are not symmetrical, and with only n=100 could be pretty easy to cherry pick to try to find something to write about).
Critical reading. Startups are such different beasts depending on what stage they're in.