The Harvard Library web team had ambitious goals for our site search, which is the centerpiece of the homepage. The goal was to bring together results from website, like building hours and staff contact information, with the online catalog, that includes books and online articles. When the site first launched, the search box included the placeholder text “Try Santo Domingo or Media Lab” to indicate the variety of results you could find through this combined search box.
However, the team received feedback via our in-site feedback survey that users were unhappy with the search experience. Based on this and anecdotal feedback from staff, we learned that the search box did not match users’ mental model of how library search works. I lead a research project to investigate these problems in more detail.
Method
I reviewed our search box analytics and conducted a concept test on an experimental version of the search box using a prototype to inform a new iteration of the search box.
This experimental version of the search box included a drop-down that appeared as you started typing your search that would allow you to jump to the online catalog results, website results, or digital collection results.
We wanted to know
- What were the most common types of search terms entered in the search box?
- Do users find a search box drop-down that allowed them to choose a specific kind of search helpful? Do they find it overwhelming?
- Do users say they’re likely to use the search box drop-down functionality?
- Are there any other usability issues with the search box?
We learned that
- Our analytics showed that over 80% of search terms were for books & articles
- Users reacted positively to the search box drop-down and rated likelihood of use very high.
- Users wanted the HOLLIS catalog to be the first option and the Library.Harvard option to be called “website.”
- 90% of participants recommended we implement the experimental search box with the drop-down instead of the current search box.
- All participants wanted more context around the search box to understand what they were searching.
Impact
The team implemented the new style search box with the drop-down, but launched it with the intent of reviewing how it was used over three months. I reviewed analytics for three months after it launched, and found that the majority of users did not select one of the search drop down options. I also saw that most users immediately selected the online catalog results from the search results page. Reviewing the analytics demonstrated that users did not actually select an option from the drop-down, even though the feature tested well in our concept testing.
After this three month review, the team changed the central search box so that it pointed directly to the online catalog, rather than the combined search results page. We also updated the search box button to say “Search HOLLIS” and added some helper text underneath the search box to highlight website content. The search button in the top-right corner of the main navigation allows users to search the website instead of the catalog.
As a follow-up to the research, I reviewed the analytics for the site search and saw a significant decrease in catalog search terms, proving that users were not mistaking the site search for the online catalog search. We also heard positive feedback from all kinds of users about this change via our feedback form and directly from users.
Amy Deschenes 