Understanding drupal.org User Behavior from Google Analytics

Shyamala's picture

Understanding user behavior using Google Analytics.
To do the same we will need to take small steps in the way drupal.org code is rendered as detailed below:

  1. Add on page js tracking, this is adding a small piece of js to the file download links throughout the site.
    Example the URL for Drupal 7 tar download should have the js embedded into code as in attached file.
    Reference:On page js tracking: http://www.google.com/support/googleanalytics/bin/answer.py?answer=55529
  2. Identify which part of the site the user visits by adopting proper url strategy. Most of our pages read node/, to help identify he same we should come up with a site URL architecture.
    eg. All documentation pages should read: http://drupal.org/documentation/page-title, forum pages:http://drupal.org/forums/forum-name/page-title and issue queues: http://drupal.org/project/issues/project-name/page-title

    This will help us identify which are the section users spend more time on, understand user behavior based on user segments,

  3. Introduce a thank you page on completion of user registration, or submission of member registration.
On Page Tracking.html219 bytes


I think you need to explain a

yoroy's picture

I think you need to explain a bit more on why we'd want this, what it will help us learn and what kind of improvement we can expect to make based on these stats. Also moving this post to the Prairie group.

There are other ways to gather download stats, too

dww's picture

Also note that there's already a large body of work underway to collect download stats (including from things like drush make, wget, etc, that would all be missed by JS-based counters on the download links on drupal.org itself). See http://drupal.org/node/32124 for more. We're actually already collecting and storing the data, it's just not exposed in the UI on d.o itself yet.

Drupal.org URL stratergy

Shyamala's picture

@dww thanks for the information on downloads. Will look this up.

Yup understand that we can not track the drush or wget downloads using JS as well. But even a partial stats on Google Analytics would be valuable considering there is a lot you can do with the Analytics. We can understand how a user got to the download link? The performance of the Download links across various countries in the world? This probably not as important as the URL strategy.

Are you aware of any reason why the no URL strategy is adopted for the drupal.org website? With out this it will be very difficult to understand where the users spend time and how they navigate.

Some of the statistics we could gather before redesign would be, understanding behavior of different segments of users. The Segments could include: Visitors, Spectators, Contributors, Larger Contributors. Only if these statistics are collected before a redesign can we quantify the benefits of the redesign itself.

Why we don't have human readable URLs on most of d.o

dww's picture

mysql> select count() from node;
| count(
) |
|   739224 |
1 row in set (0.26 sec)
mysql> select count() from url_alias;
| count(
) |
|    22278 |
1 row in set (0.01 sec)

If the 2nd number was equal to the first number, the performance of the d.o database would probably grind to a halt. :(

Performance is the key but

yarrait's picture

Performance is the key but its becomes easier to remember paths with node/$nid. Sometime(Most of the time) its become difficult to remember paths with long aliased urls.

A seemingly random number is

LewisNyman's picture

A seemingly random number is more memorable then meaningful words? Do you have any evidence to support this?

That's my personal preference

yarrait's picture

That's my personal preference I do not have any supporting evidence for it.
But take example of this post only.... http://groups.drupal.org/node/144404 is much more easier to remember then http://groups.drupal.org/redesign-group/understanding-drupal-org-user-be...

Neither of those examples

lisarex's picture

Neither of those examples seem more memorable.

The only thing we can and should do is alias the major landing pages and other important pages with something short e.g http://drupal.org/community and http://drupal.org/drupal-media-kit


That's not very efficient as

BarryS's picture

That's not very efficient as Google Analytics is used for depth-pages statistics. If we make aliases for important pages only, then we will miss Google Analytics efficiency. Due that both examples of page naming seem not so easy to remember, a global site alias definitions should be used to be able to work with Google Analytics.

bijuterii argint

Shyamala's picture

Are we not permitted to optimize the Query like what Pressflow does?

Drupal.org is already using

greggles's picture

Drupal.org is already using that.