This is my fifth post about bitly (bit.ly, bitly.com), the URL-shortening service, in the past 18 months. The first three, written in early July 2010, were focused on two simple questions: (1) how large was bitly’s pool of unique hash-string URLs; and (2) at the then-projected rate of growth, when would that pool be exhausted? Or, more to the point, what could they do to avoid running out? The respective answers were 62^6 = 56,800,235,584 (see Part I); and roughly the year 2030 (see Part II).
Part III offered some tongue-in-cheek suggestions for postponing that day of reckoning: stretching to 7-character strings; adding other character sets — Greek, Cyrillic, Hebrew, math symbols, even Dingbats; seeing if Estonia would be amenable to hosting the “characters” hu.ee, dew.ee, and lou.ee; or just waiting until nearer 2030 for the inevitable introduction of the iLobe implant with content pushed over VerizoNeural.
The fourth post led me to the belated discovery — read “public self-embarrassment” — that Twitter had introduced its own in-house shortener, t.co, at the same time as my hot air in Parts I-III. A pretty obvious reason for the drop-off in bitly traffic. That immediately changed my perspective — it was less a matter of when bitly would run out of unique hash strings and much more a matter of when the world might run out of bitly. Would the dearth become a death, not to put too fine a point on it?
Now, six months after that post, I decided to check on how bitly was doing. Unique visitors per month is an imperfect placeholder for all the data that would be nice if I was trying to flesh out an “equation of state” for social media sites. But, hey, it’s there at Compete, and it’s free!! Besides, it’s not like I’m trying to turn this into a Ph.D. dissertation. It’s just an idle curiosity that got my attention. So in that spirit, here is an update of the two long-term traffic graphs from the last post.
For combined traffic — bit.ly, bitly.com, and t.co — including the sum of all three:
And for the month-by-month history of bit.ly:
It would be easy to infer, especially in comparison to t.co in the upper figure, that bitly is in decline. Indeed, with the data since May 2010, a straight-line projection for bit.ly alone shows it zeroing-out in about 210 months. An infinitude on the technology timescale, but probably not the robust behavior that all those bitly investors would like. If we add bitly.com to the mix, however, then things look better. In fact, here is a dog-leg pattern, with the data for February through April of last year shared in both fitted lines. The correlations on both segments are very good (R ~ 0.87-0.89), suggesting that there was a pretty dramatic change in ±March 2011.
Having completely missed the advent of t.co the last time around, however, I am not going to speculate on what causal factors might have occurred a year ago to prompt this apparent up-tick.