Back to top

Drupal Download Statistics for April 2007

Background on the Numbers

The lowest level that registered in Analog this month was 113. So, if your project or release was downloaded fewer than 113 times it won't register. I say project OR release because it's possible that the 5.x-1.20 release of your module was quickly followed by the 5.x-1.21 release - the 5.x-1.20 version may not register even if the module is fairly popular. So, 1) the numbers are inherently flawed a little bit 2) project maintainers should only create new releases when there has been a significant enough change to warrant it. Doing so on every new commit to the project is a waste of your time, your users time, and reduces the quality of these metrics.

That said, on with the report...

Popular Modules

Rank Project Downloads
1 views 8,826
2 image 6,959
3 cck 6,801
4 tinymce 5,616
5 event 3,300
6 date 3,034
7 img_assist 2,834
8 gallery 2,819
9 pathauto 2,744
10 gsitemap 2,647
11 ecommerce 2,632
12 imce 2,618
13 fckeditor 2,422
14 calendar 2,369
15 widgeditor 2,280
16 acidfree 2,107
17 imagefield 2,096
18 contemplate 2,056
19 captcha 1,988
20 update_status 1,983

So, once again views takes top honors. Congrats to Earl. I included a long list today because I'm surprised that update_status isn't higher. It's a "must download" module in my opinion.

It's also interesting to note the similarities between this list and the Top 40 from Lullabot. I'm not saying the top 40 was right or wrong (I should make my own damn list!) but it is the distilled wisdom of several really knowledgeable Drupal power users. So, to me that reinforces the value of the downloads as a measure of module quality and also reinforces some of their choices.

Popular Themes

The popular themes shows the dominance of the themes that are early in the alphabet. With a few exceptions (notably zen) this reads like the beginning of the project/Themes page. Which makes me wonder if perhaps a random sort or a date ordered sort would be better.

1 abac 3399
2 aberdeen 3353
3 andreas01 2960
4 bluebreeze 2846
5 zen 2563
6 amadou 2388
7 aurora 2216
8 danger4k 2071
9 glossyblue 1899
10 barlow 1856

4.6? 4.7? 5.x?

Looking at downloads for Drupal core:

drupal-4.6.11.tar.gz 579
drupal-4.7.*.tar.gz 9,270
drupal-5.*.tar.gz 46,773
drupal-6.*.tar.gz 628

The development code that will become Drupal 6 is already more popular than Drupal 4.6. That's nice. But let's dig a little further into those 4.6 people.

Looking at the overall file downloads by the version implied in the filename:

4.7 20,279
5.0 10,525
5.x 272,193
5.x-2.x 586
HEAD 4,741

No modules for 4.6 were downloaded (or at least not more than 113 times). Which to me means that Drupal has succesfully made the 4.7 and 5.0 versions compelling enough that people are no longer holding onto 4.6, except, perhaps, for a few intranets or other similar small sites.

Similarly, it's pretty clear looking at 4.7 downloads vs. 5.x that users feel that Drupal 5 is a compelling improvement over 4.7. As the community has discussed the life/death of php4/php5 and how php5 hasn't been compelling enough to hosts for them to upgrade, it's important to make sure that each successive version of Drupal has the right features to make it more compelling so that people will upgrade. So far it seems that is the case.

Also, that's still a lot more folks downloading HEAD than I'd like to see. So, another bit of advice to module maintainers that you should branch and tag your modules.

Drupal Profiles

Finally, this is the first month where we had data about profiles. It's not a lot, and I think it's pretty heavily skewed by the fact that these are so new (e.g. dfc was the first profile, I think that's more the reason for its popularity than churches downloading it).

dfc 961
RussianInstaller 204
drupalhe 123

That's it. If you don't like my analysis, download the data and do your own ;)

Category: 
People Involved: 
timeline: 

Comments

I don't

Sadly all the data I've got here is all that I've got.

Over in a groups.drupal.org marketing post you can see a list of statistics that I feel are important.

Perhaps you could review that, agree (or disagree) and then we could split out how to gather those statistics in a repeatable manner. My feeling is that if the process is not drop-dead-simple (i.e. it runs automatically) then it is not worth working towards. We've got a couple dozen people keeping statistics about the proejct all using different methods and none of it is easily repeatable. So that it can be trended and reviewed over time. That won't work in the long run...

I discovered the marketing

I discovered the marketing group soon after finding your post :) Has automated publishing of stats been discussed before (I'm haven't followed infrastructure discussions)?

I was looking to see if drupal.org user account growth rate could be correlated with the download rate.

Lara

I discovered the marketing group soon after finding your post :) Has automated publishing of stats been discussed before (I'm haven't followed infrastructure discussions)?
I was looking to see if drupal.org user account growth rate could be correlated with the download rate.