Yet another "best time to Tweet" graph was brought to my attention today. I've spilled some ink on these in the past--they aren't uniformly bad, but they're very nearly so. Yes, there may be a time or day of the week when your tweets are replied to with greater frequency, or clicked on at higher rates. Random numbers clump; the fact that you spotted such a cluster doesn't mean a thing until you prove it. Still, these data can seem so convincing, especially when they are bundled into a slick-looking infographic. And, there may in fact be some specific situations where the success of a tweet does indeed have a temporal component. But, before you make that leap, consider this graph, printed in The Guardian earlier this year, of global cancer rates:
Observing increased activity around your tweets at certain times of the day, and then declaring those times as the "best times to tweet," is kind of like--no, it's exactly like--saying this:
Denmark causes cancer.
Kinda silly, isn't it?
Sometimes, these clusters are random, and sometimes they aren't. Do the work. Find the real reason for the clump (in the case of the Danes, the high rates are at least partially attributable to the fact that their doctors are so good that they diagnose more cases). Do that, and you'll beat the pants off the guy sending his tweets at 6 am every Sunday, six ways to Sunday.
Or, in the immortal words of Brian Fantana, 60% of the time, it works every time.