Drupal download statistics

Last year around this time, I shared the download statistics of Drupal core so I figured that an update was in order. It looks like last year, from May 2006 to April 2007, Drupal core was downloaded more than 600,000 times.

Absolute download statistics
Relative download statistics

These numbers do no include betas, release candidates or CVS checkouts. Also, we can't track downloads from mirrors, such as various Linux distributions, nor can we track installations through control panel software for hosting like cPanel or Plesk. Contributed themes or modules are not included in these numbers: we only looked at the main Drupal download.


Jelle Druyts (not verified):

Interesting figures...

What I find peculiar in the bottom graph is the fact that even when a certain release has been superseded by another one, there is still a "bump" in the old release downloads after a while. For example in December '05, the percentage of Drupal 4.5 downloads goes up for a moment even though 4.6 has been released for a while and 4.5's download share was stable before that (going down slightly or flatlining at best). Same around November '06 for 4.6 and now again in April '07 for Drupal 4.7. This seems to be a trend.

But that would mean that users are "abandoning" a newer version in favor of an older version. Maybe I'm missing something here, but I'd be a bit worried :-)

Also, the impact of Drupal 4.7 was much stronger than that of 5.0: where 4.6 dropped down significantly by about 90% when 4.7 was released, Drupal 4.7 only dropped about 60% towards the "major" 5.0 version when that came out. I'm not a Drupal dev, but if I remember correctly there were some larger breaking changes in 5.0 that might have caused this.

Anyway, I'm not an analyst so it would be nice to get your interpretation of these figures :-)

Gábor Hojtsy (not verified):

Put on the dates of the bugfix releases of earlier versions on the graph, and you might immediately get an answer. (I did not do it, but it is a reasonable educated guess). Bugfix releases are rolled even after a new major version comes out, and they *should* bump interest in earlier versions.

Eivind (not verified):

It's reasonable that each new major release will see slower and slower uptake.

Think about it. When the project was young, few people had much invested in it, few people where running huge sites on it. And so most users where forefront-runners, bleeding-edge lovers. People who *always* get the newest version.

As Drupal matures though, you're going to have a lot of older sites running stable websites on a certain version of Drupal. Many of those subscribe to "don't fix what ain't broken", instead they download the latest bugfix-release of the version they already have. "new and exciting" is a *bug* to them, they don't want exciting, they want *rock-solid*.

I personally think this is a *good* sign.

Jacques (not verified):

I like your comment, Jelle Druyts.

As I understand the issue, it is because modules are not compatible with the new release. So a webmaster who needs to create a new web site might opt for an older release that has a lot more compatible modules.

As for me, I still have Drupal 4.7.6 websites that are using modules not fully-compliant with Drupal 5.1.

Sean Robertson (not verified):

Any way to add CVS checkouts to this? We do all of our new site installs as CVS checkouts, so all of our installs get missed. I suspect a lot of other developers do the same thing (makes updates a lot easier among other things).

Tom Geller (not verified):

Any chance of running these statistics again for a more up-to-date view? If this has already been done (here or elsewhere), please point me at the URL. :) Dank je wel!