User-centred design starts with an unbiased study of user needs and behaviour. In this case study, we show how it can be employed to ensure that a web site meets the needs of its users. For Cancer Research UK, we took the unusual (for usability design) step of comparing different site designs, to help maximise the incremental benefit gained from the re-designed site.
Why did Cancer Research UK approach Minervation?
Cancer Research UK approached us in November 2007 because they needed help reorganising and re-designing their CancerHelp UK web site.
CancerHelp is the most visited of Cancer Research UK’s websites, with c.800k visits and over 2 million page views each month. The site has:
- over 6,000 pages of content
- detailed information on 42 specific cancers (including information on relevant subtypes)
- further information on 20 other types of cancer
- comprehensive information on coping with cancer and other aspects of living with the disease
- a question and answer section and
- a clinical trials database
Cancer Research UK were concerned because the content on the site had out-grown the original navigational structure making it difficult to manage, and less easy for the user to navigate. They also wanted to make sure that the re-designed site was “future-proofed”.
They put together a redevelopment project to review the site Information Architecture (IA), re-design the site, and move it to a new hosting platform. Throughout, Minervation performed as an integral part of this team.
How did we use our information science expertise and usability testing skills to help Cancer Research UK?
Our work for Cancer Research UK was divided into five phases:
1. Information architecture (December 2007)
- We conducted an independent review of the site information architecture (IA)
- We produced a report with recommendations for a new IA
- We presented these findings to Cancer Research UK
2. Usability testing 1 (January 2008)
- We produced a “wireframe” prototype website (i.e. the site did not have any graphic design, just text) using the proposed new IA
- To ensure comparability, the existing site was also rendered as a wireframe
- We tested these wireframes for usability with a representative sample of participants (patients, carers, health professionals), comparing the existing web site with the proposed IA
- Participants were asked to complete a series of tasks, such as finding information about treating advanced bladder cancer
- We recorded the time taken to complete tasks, completion rate, errors made and qualitative feedback from participants
- We presented these findings to Cancer Research UK
3. Usability testing 2 (April 2008)
- We ran further usability tests (with a separate representative sample of users) on a fully designed prototype
- This time the prototype we used for the test looked pretty much like the finished site, with graphics and the new proposed site layout
- We ran the sessions as before except that we left the participants to choose the specific questions they were interested in answering. For example, we asked people to put themselves in the position of someone interested in finding out more about bladder cancer, but we allowed them to generate their own questions and then look for answers on the prototype site
- Again, we recorded quantitative and qualitative data on task success
- We presented these findings to Cancer Research UK and they used them to redesign and rebuild their website, which was launched in October 2009
4. Usability testing 3 (January 2010)
- We ran further usability tests (with a separate representative sample of users) on the relaunched website
- We ran the sessions as we had in April 2008 so we could compare the results
- We found that 90% of questions generated by test participants could be answered within 3 clicks. The IA we have built with the Cancer Research UK team allows users to quickly and easily answer their questions
- We discovered a variety of small but notable points for development and fed these back to Cancer Research UK who adapted their site accordingly
5. Usability testing for Cancer Chat (March 2010)
- We ran further usability tests (with a small sample of users) on the online discussion forum part of the Cancer Research UK site (Cancer Chat)
- We made a number of recommendations for change, which informed the relaunch of the forum.
How did Cancer Research UK benefit from this work?
“I have just looked at the new usability report and analysis and it’s great. It’s very useful to have numerical data comparing the effectiveness of both information architectures. Thanks!”
– Liz Woolf, Head of CancerHelp UK
- They received an independent evaluation of their existing web site, with practical and achievable recommendations for improving the site architecture and design to better meet the needs of their audience.
- We worked closely with their team to build prototypes for testing that investigated the large number of issues under scrutiny.
- We helped them simplify their site architecture from 12 top level categories down to 4; a change that has greatly improved the usability of the site. Comparing the results from our tests in 2008 and 2010 we have shown a 19% improvement in the information architecture.
- We are working closely with their team to ensure that the usability testing service they receive is focused and appropriate.
- We recruited test participants and ran the sessions in a flexible, creative and sensitive manner.
- We captured the views of the participants and presented them to the Cancer Research team in a variety of formats (written report, face-to-face presentation, DVD of test clips) for maximum impact.
- We supported the Cancer Research team in making a number of difficult decisions about the future direction of their web site.