Retention of early Slashdot users

Still some around

With the recent resignation of Rob "CmdrTaco" Malda as Slashdot editor after 14 years at the helm, a number of early Slashdotters came out of the woodwork to express their appreciation or comment on some aspect of the site's history.

This made me wonder: just how many early Slashdot users are still active on the site, anyway? The answer was higher than I expected, and not only due to long-gone users who returned yesterday to comment on Malda's departure.

I scraped the Slashdot userpages of anyone with a 3-digit or lower UID, which covers the people who signed up in the first few days after user accounts were implemented (in 1997). Slashdot had already existed for a bit at that point, but usernames previously were entered on a per-comment basis, like on a Wordpress blog, with proper account registration only coming later. I chose 3-digit UIDs as the cutoff because that's how long mine is, of course.

Of those 999 UIDs, 310 of them either: don't have accounts associated with them (or don't any longer, anyway), have never posted a comment, or must've posted their last comment in 1997 or 1998—Slashdot userpages only include comments from 1999 and later, due to a software change, so I can't say for sure, at least not without much more extensive scraping.

That leaves 689 early users who exist and have posted at least one comment between 1999 and today. Of those, 179 are still around, defined as having posted at least one comment in 2011. Did they all return just for resignation-of-Taco day? Twenty did. But 159 had been active in 2011 for reasons other than Malda resignations. That's actually higher than I expected; 16% of the 3-digit-or-less UID space, and 23% of those active since 1999. A quarter retention over more than a decade ain't bad. Looking at the distribution of most recent comments, the dropoff appears to be roughly linear, with 30–50 early users permanently leaving per year. If the linear trend continues, the last 3-digit Slashdotter will stop posting around 2015 or 2016.

Here's the overall retention graph, with one line using users' most recent comments as of today, and the other using them as of August 24, 2011, prior to the 20 early users returning to post about Taco's resignation (so e.g. if their last post before that had been in 2008, as with Hemos, they count as having dropped off after 2008):