Library User Experience Research blog: Welcome!

Update: When less is more in a discovery service UI

Back in October I wrote about an analysis we conducted on measuring the use of point-of-need library help within our discovery service interface. That analysis was based on 3 months of data, and you can see a visual summary of the initial findings from my ERMN conference session slides from last Fall.

It's February, and I've rerun the numbers with 6 full months of data to see if anything had changed. With the ebb and flow of courses, it's reasonable to think that use patterns could change significantly as students get deeper into their courses and library research.

Final results:

Overall, the results have stayed mostly the same. Some of the charts I shared previously were based on pageviews. I've updated these because it's a more accurate reflection to show the data by sessions with Google Analytics campaign data. I've briefly summarized the updated findings, and I recommend reading the original post for context and commentary.

• A help link is clicked in 1.7% of EDS sessions. This aligns exactly with the number of students reporting on a recent post-evaluation course survey that they used library help during their course. That 1.7% still translates to a whopping 16.5K users, 40.1K sessions, and 60.1K pageviews.
• A help link located in the top menu is clicked by 7 times as many users and in 17 times more sessions than a help link in the right sidebar (only 0.1% of sessions). Within this top menu, help was viewed 3.5 times per user for an average of 4 minutes. I'll call out again that our "Databases A-Z" list is the bulk of this activity, which is arguably not library help per say, but we're taking a broad look at how our users are interacting with supplemental library resources that take them outside the discovery service.
  
  
  
  
  
  
   
• 34% of users who click "How to Search" leave after visiting this page. In other words, at least a third of all users are only (maybe) reading the guide's introduction, then leaving. This tells me that they likely didn't find what they were looking for immediately.
• Students who click "Ask a Librarian" in the EDS menu only submit a chat or email 6% and 7% of time, respectively.
• Quick Answers (FAQs) in the sidebar are clicked by over 2 times as many users and sessions than guides in the sidebar.
• Quick Answers in the sidebar are not listed in order of popularity (most to least viewed). The first Quick Answer listed (How do I search Thoreau?) is 2nd to last in popularity.
  
  
  
  
   
• The guide rating mechanism to measure the usefulness of library help when coming from EDS was mostly a failure. Only 13 users shared with us whether the help was very/somewhat/not useful. The method of capturing user feedback is likely not visible enough to be effective.
  
  
   
• For the enhanced "no results" help pilot in the Education EDS search profile, the video is clearly preferred over the other linked help. Users viewed the video 2.3 times over the other no results help options combined. Users also watched an average of 81% of the video.
   

What this boils down to is:

1. Users access help embedded within EDS at about the same rate they report accessing library help generally. They rarely access library help regardless of where it's located. This shouldn't come as a surprise because we already know that students come to the library first-and-foremost to conduct research, not to build their library skills :p. If they get stuck, then they might seek help (or give up).
2. But when they do access library help in EDS, they're primarily using the resources available in the highly visible top menu navigation.

Recommendations:

Based on these findings, here are the recommended adjustments to the to our EDS interface to provide more effective point-of-help without distracting from the research experience:

• Top navigation:  Refine menu options and increase visibility and usability. Specifically--
  • Add icon to "Ask a Librarian" link --> align with website and visually call it out
  • Update "Thoreau: New Search" to "Start new search" --> simplify language
  • Remove "Get Library Help" --> low use
  • Increase font size and add external link icons to left-justified menu items --> increase visibility and inform user they're leaving EDS
  • Focus and clarify the "How to search" link --> A couple recommendations to consider here: Students are already in EDS and trying to construct their searches using keywords. So let's link them directly to Set Up Your Search page. They can always visit other pages within the guide, but I'd be willing to bet most are struggling to construct their searches properly to get optimal results. Another option is to link to the How to search Thoreau page and make some improvements to the content here.
     
• Sidebar:  Remove this altogether. It's ineffective and takes up valuable real-estate. Moving to a 2-column layout will reduce visual clutter and the number of choices students have to make (cognitive load), and it would put more focus on the search filters and search results.
  
  
   
• No results:  Implement the enhanced "no results" message across all EDS profiles with a few adjustments--
  • Develop a ~1min 30sec video focused on setting up an EDS search successfully that works for all EDS search profiles (to replace the Education-focused video)
  • Update the "Techniques for searching library databases" link with the Set up your search page link --> direct access to the help they need in this scenario
  • Remove the Ask a Librarian link --> it's just not clicked
    
    
     
• Guide ratings:  Remove. Not an effective way to measure usefulness of the help content.
   
• Other considerations:
  • For better visibility to targeted, topic/subject-based instruction, let's consider moving the Springshare "Recommended Research Guides" placard from the bottom to the top of the search results (screenshot above). Do we risk annoying lots of students and faculty by pushing results further down the page?
  • Long term: Explore integrating software that can walk students and faculty through a short, interactive demo within the actual EDS interface (e.g. WalkMe or Userpilot).

Of course, these recommendations will be shared with library staff with the opportunity to provide feedback. Usage data analysis, if done properly, is one excellent way to assess web content and design and to identify glaring issues, but it rarely provides the full picture, the "why", or the nuance of usability testing. For this project, we won't be able to do usability testing to further inform these recommendations (there just isn't time or resources to test every change we make to our systems), so we'll plan to continue measuring and making iterative adjustments with the goal of providing a stellar research experience.