Drupal.org – what we learned from the card sort

If you’ve been following this project you’ll know that we’ve been doing an online card sort recently to help inform the information architecture for drupal.org. To date we have had more than 200 people participate in this card sort exercise, which is a tremendous effort and a bucketload of data! Thank you all!

In particular we were interested in understanding how we group and label different types of content, and whether the language (as in terminology) varied between insiders and outsiders.

The card sort is still running but we have done an initial analysis of the results to date – the findings are not really so surprising, but nonetheless, useful to have available.

Essentially, what we found is that insiders and outsiders generally use similar words to group and label content where that content is not particularly specific to Drupal, or doesn’t involve ‘Drupal-Speak’

So, for example, for both insiders and outsiders there were lots of occasions where a significant majority of users grouped terms into the same category, for example:

  • ‘Local User Groups’ was grouped into ‘Community’
  • ‘Installation Documentation’ was grouped into ‘Documentation’
  • ‘General Concepts’ was grouped into ‘Getting Started’
  • ‘Getting started’ was grouped into ‘Beginner/New/Getting Started’ and ‘Documentation’ – both equally split in this way in both groups!
  • ‘Features and Mission of the Drupal Project’ was grouped into ‘About’
  • ‘Administer Drupal’ was grouped into ‘Documentation’

However, when it came to content items or terminology that was particular to (or used in a particular way within) the Drupal community, the split became  more apparent. Drupal ‘insiders’ familiarity with the current structure of the drupal.org website also seemed to come into play – in fact, sometimes people actually responded with a specific URL! (eg. groups.drupal.org). Examples of these kinds of content include:

  • Contributed Modules – responses from ‘outsiders’ were all over the place and didn’t show much of a trend at all. Terms included community, customise, design, development, features, modules, enhancing Drupal, understanding Drupal, getting involved, mastering and  more. On the other hand, ‘Insiders’ showed two very strong categories – Download (where modules reside in the current IA) and Developer
  • Core Project – most ‘outsiders’ put this in a category called ‘About Drupal’, where as ‘insiders’ unsurprisingly used the same cateogories as for Contributed Modules (both are considered ‘downloads’ in the current IA
  • Themes – again, the categories suggested by ‘outsiders’ was wide ranging, but most commonly suggested was ‘Customise’ and ‘Themes’, were as by FAR the most common category suggested by ‘insiders’ was ‘Downloads’.
  • Same for Translations, which clearly went into a ‘Downloads’ category for ‘insiders’ but left ‘outsiders’ perplexed – again, no clear trend emerged from them and suggested categories ranged from Customise, to Advanced Help to Projects.

You  may also have noticed that, in general, suggested categories were quite broad – there was extensive use made of the categories ‘About’, ‘Community’, ‘Documentation’, ‘Developing/Developers/Development’, ‘Get Involved’.

So, what does this mean for our project?

A card sort is never intended to ‘set’ the information architecture, but is rather used as a ‘probe’ into the existing and potential audience to get a sense of how they make sense of the content that you are trying to organise. We are not going to take the labels suggested by most participants and map the content to those labels and call that our IA, but we have learned some very valuable lessons. Including:

  • We need to be very careful and aware of Drupal-speak – it causes no end of confusion for people who are not familiar with Drupal. This doesn’t mean that we abandon it – after all, it is part of the efficiency of communicating within the community. But we need to make sure that we don’t use it for major ‘sign posts’ in the information architecture and that when we do use it, we add ‘supports’ for new players (outsiders)
  • Opportunities for cross-referencing – there were several instances where a piece of content showed more than one ‘trending’ category, for example the inclusion of ‘getting started’ content in both a ‘beginners/getting started’ category as well as a ‘documentation’ category. There are several instances like this where we can ensure that content is cross-referenced from one section (or sub-site) to another based on expectations shown in the card sort data.
  • Category naming – for example, a term like ‘Community’ is not currently represented on the Drupal.org website but are widely expected by both ‘insiders’ and ‘outsiders’ alike. This supports the consolidation of the more ‘social’ aspects of Drupal (groups, forums etc.) into a major section labeled ‘Community’.

Going forward

We will use the insight we have gained to date from this card sort to help inform the proposed Information Architecture for the Drupal.org website, which of course we will continue to share with you as we proceed.

It is *possible* that we may conduct a further card sort in the near future, however this one will be a ‘closed’ sort, where you will be asked to place content into a set of pre-defined categories. I’m only half convinced that this is a good approach (having had mixed success with closed card sorts in the past… ), so I’d be interested in any thoughts you have about whether this would be an interesting and useful exercise.

I’m glad to hear that several of you really enjoyed participating in this exercise and I do thank you for taking the time to do so.

Drupal.org – Initial thoughts on the Information Architecture

Firstly, I want to start by thanking you all for your participation to date in this project – it has been beyond helpful! Of course, now I’m going to ask for more, as we move forward to working out what the Information Architecture for the new Drupal.org is going to be.

If you have been following this process you will know that the overwhelming response to the question of ‘one site or many’ has been – many. And for many very good reasons, covering issues from technical and implementation through to community, scalability and usability. So, it should be no surprise to hear that we’ll be proposing to take that approach to the information architecture.

In terms of how it breaks down, here are some initial thoughts on which I’d value your feedback.


We propose that Drupal.org have, in effect, two homepages. On ‘not logged in’ and one ‘logged in’.

The ‘not logged in’ homepage would serve as the public face of Drupal to ‘outsiders’ (those with little or know knowledge of Drupal) and will be the place where we can introduce Drupal and ‘sell it in’ – showing it’s features, capabilities, reliability, the scale of it’s use across industries and around the world, and it’s amazingly active community.

We will encourage people to ‘join’ Drupal.org so that they can access the ‘logged in’ version of the homepage. The purpose of the ‘log in’ is not to ‘hide’ content, but rather to ‘activate’ the homepage as a tool for managing the vast amounts of content and activity that make up Drupal.org (or d.o as the ‘insiders’ call it!)

We imagine the ‘logged in’ version of d.o to be similar in approach to Netvibes – the idea being that there are a range of ‘widgets’ available that you can select and arrange as best suits your needs – so depending on what you are most interested and engaged in, you can ‘tailor’ the content on your homepage to best reflect this. Developers might also be able to develop new ‘widgets’ and submit them for use on the d.o logged in homepage.

We believe that a single login solution is imperative (and we’d like to suggest that supporting OpenID would be an excellent idea).

We also advocate a unified search across all of the d.o content, with a search results page that better identifies the type of content found and it’s source, and good filtering capabilities.

‘Portal/Network’ Header

Although we agree that the ‘many site’ approach is the correct approach for this project, it is important that we resolve a current issue with the information architecture which is that many of the current sub-sites are unfindable – we need to make these visible and easily accessible to all.

The logical way to approach this is to use a ‘network header’ which is global – shown on all pages of all d.o sites – and which contains links to each of the sub-sites in the d.o network.

A different visual treatment may be developed for some or all of the subsites to allow for their specific purpose and character, whilst clearly positioning them as a part of the Drupal online presence. We want to balance the need for flexibility and scalability with their inherant threat to usability.

A few people mentioned jquery.com as a reference site and – broadly speaking, in terms of the header/sub-site treatment, this is a guide to the approach (not to any great level of detail and not from a visual design perspective of course!)

It is quite a big task to decide what ‘sections’ are required and what requires it’s own ‘sub-site’, but here is a first run at what we are thinking.

Sections of d.o (but not their own separate subsites):

  • About (includes marketing information – features, benefits, demo etc. – also information about Open Source in general)
  • Jobs (pulling it out of the ‘groups’ site and making it more prominent)
  • Events (also pulling it out of the ‘groups’ site and making it more prominent)
  • News (similar content stream to current d.o homepage. Include link (at least) to Planet Drupal here)
  • Get Started (a quick start guide for new players, helping to get them over the hurdle of getting their Drupal site up and running and showing them how to access the community and other help)
  • Get Involved (an overview of ways that people can get involved in the Drupal project and/or community)
  • Professional Services (a directory of hosting, design and development services for rent, possibly paid listings?)

Sub-sites of d.o

  • Community – the social hub of d.o hosting groups, discussion forums etc.
  • Documentation- where all drupal related documentation is hosted, including Handbook and API
  • The Drupal Project – the project management aspect of the site for core, modules, themes etc. Issue lists etc. live here
  • Download – essentially an alternate ‘window’ on The Drupal Project, but focussed on helping people locate then download core, modules, themes etc. (rather than the ‘development’ aspect of them)
  • Association – as per the existing Association content

Feedback required:

As I said, this is the first run at the structure and I’m looking for your feedback on where we need to make improvements. Specifically feedback that I’m looking for is:

  1. Is something missing? Is there an important part of d.o content that you don’t see fitting into one of these categories?
  2. Technical issues. Is something that I’m suggesting here going to cause more technical pain (which will equate to implementation delay) than it’s worth? Let me know.
  3. Drupal Project / Download – two windows on the same content – what do you think of this idea? I’m not sure that it’s entirely right, but we *do* need to resolve the fact that people who access modules just to ‘download’ them seem to have entirely different needs to people who access them to ‘develop’ them.
  4. Design and UI – perhaps this is just my personal mission (I hope not!) but I’d like to make an information architecture that was more appealing to people like me who might want to get involved with the project. I was toying with creating a section specifically for design/UI/UX… but it didn’t feel quite right. At the moment I’m imagining it living under The Drupal Project… what do you think?

I have some images that I’ll put up shortly to roughly illustrate these for those who prefer pictures to words (don’t we all!) – my internet access doesn’t allow me access to my images at the moment… grr.. but I thought I’d get this up now rather than wait.

Looking forward to your thoughts on this!

Updated: here are a couple of images that might illustrate my thoughts a little. Larger versions available on Flickr (via links below)
Larger version available here.

Larget version available here. (Note: this is not the proposed design for the header, just a sketch of all the ‘bits’ that might be in it).

Drupal.org – One site or many?

One of the information architecture questions we need to resolve for Drupal.org is whether we try to make it one cohesive website experience, or whether (as it is now) there is a ‘family’ of mini-sites that make up the online experience of Drupal. 

The current family consists of drupal.org, groups.drupal.org, api.drupal.org, association.drupal.org and more (including a proposed developer.drupal.org).

It seems to me that it has been mostly through the organic growth of the site and of Drupal itself, some technical and perhaps some social issues that have lead to the propagation of all these mini sites. Creating a new site helped get the required content or functionality up and running where as, if it had to go into the Drupal.org infrastructure it may never have come into existence.

Some of the mini-sites theoretically appeal to all users (for example, the ‘groups’ should have almost universal appeal. Others, like the API, are perhaps more targeted towards the technical audience.

The question now is – given that we essentially have a blank sheet – should we attempt to pull everything back in under the umbrella of a single ‘drupal.org’ entity?

The proposed Information Architecture that we received as a part of the RFP for this project seemed to indicate that this was the preferred approach (ref: http://groups.drupal.org/node/10003)

In my experience, however, a better experience is usually created by trying to architect content into one website, rather than a family of sites. There are, obviously, exceptions to this!

I’m interested in hearing what you think about this question? What do you think are the advantages and disadvantages of the ‘one site’ approach? What, if any, content should be shelved onto a separate site, and why?

Drupal.org – Help Overhaul the Information Architecture – participate in our online card sort!

Continuing in the community led approach to the experience design, let’s get started taking a look at the information architecture.

A really useful exercise to help understand how people organise, understand and label/name information is to do a card sort. (Ref: A definitive guide to card sorting)

As we’re scattered all over the globe, we’ll have to settle for an online version. I’ve actually set up two versions of the card sort because I particularly want to understand differences between the way that ‘insiders’ and ‘outsiders’ deal with drupal.org information.

If you identify as an ‘insider‘ (which might reflect either your expertise and/or closeness to the drupal community) and you want to participate, please go to this link and participate in the cardsorting exercise.

If you identify as an ‘outsider’ (particularly if you’re new to Drupal, but also if you don’t feel close to the community) and you’d like to paticipate, then your cardsorting exercise can be found here.

It should take you about 15-20 minutes to complete the exercise, so be sure to get yourself a nice cup of tea before you get started!

This is just an information gathering exercise – it’s not going to define the Information Architecture, just help give us the understanding we need to help shape it correctly. Don’t worry too much if there is information missing (we’re just using a sample of the entire site(s)) and don’t worry if you don’t understand what a card means, and feel free to leave copious comments as you go through the exercise – there is a space for you to do so.

As ever, I’ll let you know what interesting things we learn as they come to hand.

(Remember, you can also participate by submitting some wireframes!)