Another UX research project has begun! NYU Libraries has a 24/7 chat service for its users called Ask a Librarian (AAL), where library staff answers questions from user, similar to the chats you see on e-commerce websites. AAL is a very popular service among users, getting about 1500-1900 chat requests per month during the semester, and is easily accessible from most library web pages through a chat window.
The range of questions asked on this platform are diverse, with users asking in depth research questions to locational questions. Without doubt, this tool is a goldmine for qualitative and quantitative research, and the UX team has been looking at capitalizing on the valuable user input to improve services. We have already benefitted from this data by creating user personas on analyzing and clustering transcripts.
In our current project, we’re analyzing a random sample of anonymized chat transcripts to identify top user concerns, ambiguous language usage, highlight seasonal needs of users, and ensure effective FAQs. We also plan to periodically assess the health of our website through chat analysis, apart from the monitoring the analytics reports. We also hope to write an evidence-based research article about this project. In the meantime, I’ll be adding monthly installments of our work on the blog.
Some immediate benefits of analyzing chat transcripts as follows:
Identify user needs.
Users voluntarily indicate areas that are unclear through their questions. This helps the web-design team to understand the frequently-used services/information and inform design decisions about providing easy accessibility to these services. We are going to randomly sample chat transcripts from the AAL software and identify recurring issues.
Reduce discrepancies in language usage.
Technical jargon and acronyms may be difficult to understand to users. For example, users may be unaware of terms such as databases and ILL (Inter-Library Loan). We will analyze transcripts to abstract user-generated language to make service names intuitive.
AAL chats can give us insights about frequently-asked questions about a service, which can be made easily accessible by mentioning them on the homepage. It will also help in devising strategies to promote various services on homepage as per varying needs of users at different times of the year. For example, during the start of the academic year, it may be helpful to flash shortcuts to “how to” sections on the homepage due to a large number of novice users.
Get ready-to-use, conversational answers for your “How to” sections.
Users frequently ask instructional questions such as “how to search for a journal article” in spite of a FAQ section. In such cases, successful instances of a librarian walking the users through the process can be used as answers in the “how to” section, lending us conversational, easy-to-understand, and humanizing responses. Good examples of process walk-throughs will be recorded to add to “how to”/FAQ section.
Improvements in web-design informed by research can be validated by analyzing chat transcripts. Reduction in the questions pertaining to the top identified issues will validate the changes in web-design. Periodical analysis of the chat transcripts can lead to a cyclic process of identifying user needs and improving the web-design accordingly.
This project is in currently progress. Stay tuned to learn about our findings! Share your comments, insights or other ways of using chat transcripts to improve user experience in the box below.