“How can we improve how users get help online? Wouldn’t it be cool if users could just click on something in an application and see questions and answers?”

That’s the question that motivated Parmit Chilana, a doctoral candidate, and Assistant Professor Amy Ko, at the University of Washington Information School, to develop LemonAid, a software program that can be added to any interactive web application to give users instant access to helpful answers when they have a problem. The user simply clicks on a label or image they believe is most relevant to their problem and they can see the questions others have asked along with the most popular answers.

“Help is a universal problem, especially when you are interacting with a new application for the first time…all of us need some kind of help and guidance,” noted Chilana.

Her findings from a large field study of LemonAid, to be published as part of the Conference on Human Factors in Computing Systems (CHI) this spring, indicate that over 70% of users found LemonAid to be helpful and intuitive and, when given the choice, would use it again. These results are a strong contrast to prior work that has shown that users are frustrated with help systems and often fear clicking on help. In addition, software development teams were easily able to integrate LemonAid into their web sites and found that it helped them identify their users’ popular help questions.

This idea of crowd-sourced help isn’t new and is in fact common in development communities, but it is often difficult to find relevant answers. Developers post questions to their peers about all kinds of issues, from coding to usability and design. End-users might find application-specific forums using a search-engine, but may not understand the technical language or references once they find them. Many end-users don’t even think to do a search when they face a problem and as a result, become frustrated with their experience.
LemonAid is a way to harness the knowledge of the larger end-user and development community to benefit everyone using an application. It is generic enough to work with any application and has the potential of saving companies’ money by transferring general and more common help questions to the broader user community to answer. Chilana presented the description of LemonAid’s underlying system and retrieval approach at the CHI 2012 conference last year.

Chilana conducted a lot of research prior to developing LemonAid.  She worked with Autodesk Research in Canada, who have 10 million users of AutoCAD and other design tools, asking their support teams how they handled requests, and found out how hard it was for the support teams to be as responsive as they would like.

She submitted and received one of five inaugural Facebook Fellowships in 2010 and was also offered a paid internship at Facebook headquarters where she developed a case study on their design process. Facebook users now number over a billion and Chilana observed that their designers strive to create a system that is user-friendly, but “good design can take you a long way, but it is just impossible given the diversity and numbers of users to always get the design right. No matter how good the design, help systems and support will remain critical areas for the future.”

Chilana tested the beta version of LemonAid on four websites at the University of Washington hosted by the UW Libraries, UW Medical Center, and the UW Information School. These partnerships allowed her to expose LemonAid more widely and gather data from thousands of visitors to those sites. This study was supported by a Google Research Awarded to Professor Ko and Chilana in summer 2012.

From the beta-testing, she was pleased to learn that LemonAid successfully facilitated support for end-users to commonly asked questions and relieved some of the burden on tech support, but she also found an unanticipated use – the question and answer format enabled users to learn more about an application itself. Even if they didn’t need help, many people used LemonAid to browse popular questions and anticipate what a feature in an application might do and whether it would be useful to them.

Chilana gives credit to her faculty advisors, Ko and associate professor Jacob Wobbrock for their advising on this project. She is quick to add that her work could not have been done anywhere but at an Information School, where researchers bring in so many different perspectives.

The development of LemonAid involved a variety of disciplines including human-computer interaction design, information seeking and retrieval principals, and knowledge of computer algorithms. She hopes to bring this interdisciplinary approach to her next position at a university or in industry when she graduates this summer.

“The iSchool creates space for interdisciplinary researchers to come in and have this playground. I feel fortunate that we can put people first when designing these types of systems rather than focusing only on the novelty of the system or the cool technical contribution. Here it is about understanding people and designing systems around that.”