NASIG 2012: Discovery on a budget: Improved searching without a web-scale discovery product

Speakers: Lynn Fields & Chris Bulock, Southern Illinois University Edwardsville

They have a link resolver, database list, and A-Z journal list. They formed a task force a few years ago to redo the webpage. They decided to approach it from a UX perspective, rather than library committee perspective.

Their initial survey of users found what most of us have learned about our websites: too many links, too much text, too much library jargon. They implemented many changes based on this, and then then did some observational study of users doing two specific tasks. This also resulted in identifying confusing aspects of the library site, so they made more changes and did another observational study. For that one, they divided the participants into two groups in order to determine which aspect of the modifications was more effective.

They took what they learned from the website studies and applied that to a study of the catalog use. They wanted to know if users could find an ebook in the catalog (distinguishing it from a print book), understand the catalog displays (and use faceting), and understand the consortia catalog interface.

Lessons learned:
There’s a gap between freshman and senior instruction. They need to develop more instruction sessions on specific topics like ebooks and facets.

Discovery is more than the journey from search box to full text, and there are many factors that impact the end result. This includes the look of a database/catalog, names and labels for resources, placement of the search box, etc.

Core lessons:
1. names and language
Would students know how to define a database, periodical, e-resource, or even research? Using action-based language is more effective. Cutting down on vendor branding helps, too.

2. order matters
First impressions are important, so arrange the order of things by importance/relevance. Minimize reading. Descriptions or lengthy cues are often ignored.

3. be familiar
If you do a search in Google, Amazon, or WorldCat, you get very similar looking search results pages. If you put important content on the right column of your search results page, it won’t be as visible because users are used to ignoring advertising on that part of the screen.

4. let the users help you
Surveys, focus groups, observation studies… and get more than just the vocal minority. Observational studies are more about what they actually do than what they say they do, and are much more valuable in that way. Capture data about errors by making it easy for users to contact you (i.e. EZproxy host errors).

5. search boxes
No box can search everything, but people will use it for anything. If the search box is limited, make that clear. Searches of database titles (not content) can be problematic, as users expect it to search for article-level content.

6. work together
Discovery doesn’t respect department divisions. Work together from the beginning.

Don’t think that doing this once is enough. Keep it ongoing, and remember that the discovery process has many steps.

tl;dr — Usability studies and improving the user experience is hard, but necessary. Discovery isn’t just about buying some massive central index of content.