Small cues that display a user’s transaction history may help a website feel almost as interactive as chatting with an online customer service agent, paving the way for more cost-effective websites, according to researchers.
“One of the challenges with online interactivity is trying to imbue the site with a sense of contingency — the back-and-forth feel of a real conversation,” said S. Shyam Sundar, Distinguished Professor of Communications and co-director of the Media Effects Research Laboratory. “What we found is that providing some information about a user’s interaction history on the site can offer a sense of conversation.”
Interactivity is natural in both actual face-to-face and video streamed conversations, but developers have found it challenging to foster this type of interaction in static websites, according to Sundar. In traditional conversations, a message is usually met with a response, creating a thread of interconnected messages.
The researchers, who report their findings online and in a forthcoming issue of Communication Research, found that visitors to an online movie search site considered a version of the site that displayed recent interaction history more engaging and said the site fostered better dialogue than sites with no or limited history. As long as this interactivity was strong enough, the presence of a chat function with a live person or chatbot did not add appreciably to user experience of the site.
Chatbots, which are software programs designed to answer questions and mimic conversations with humans, are becoming more popular on mobile devices and websites. One example is the Siri program on Apple iPhones that can answer a user’s verbal questions.
“The addition of a chat agent increases the perception of interactivity,” Sundar said. “But it does not increase the sense of dialogue or user engagement when the site already offers detailed interaction history.”
Study participants also did not find a human chat agent to be any better than a chatbot, as long as interaction history was available, according to Sundar, who worked with Saraswathi Bellur, an assistant professor of communication, University of Connecticut; Jeeyun Oh, assistant professor of communication, Robert Morris University; Hyang-Sook Kim, an assistant professor of communication and media studies, St. Norbert University, and Haiyan Jia, a post-doctoral scholar in information sciences and technology, Penn State, all former doctoral students in mass communications at Penn State
The study could lead to more cost-effective solutions for businesses that want to make their sites more engaging, but have limited funds to program a chatbot or hire a human customer service employee.
The researchers asked 110 participants, who had an average age of 21, to use an online movie store called Movie Hub to help them choose a movie to rent. They could browse, use the search engine, or ask an assistant, which was either an online chatbot or a human assistant. To carry out the assignment, participants were assigned one of five different versions of the site, ranging from a low-interaction sites without interaction history to high-interaction sites with interaction history and the ability to chat with a human agent.
Participants who interacted with the agent or the chatbot said the site was more interactive than the subjects who used the site without the chat feature. But, this did not translate into better attitudes toward the site.
“Instead, interaction history seems to foster positive attitudes by giving users the impression of a conversation,” said Sundar.
Providing this type of interaction history is both a goodwill gesture and helps to improve the user’s trust in the site, he added.
“Users appreciate the full disclosure, in part because it increases their awareness of their own actions,” said Sundar. “This is quite important in the current climate of heightened concern for online privacy.”
The above story is based on materials provided by Penn State. The original article was written by Matt Swayne. Note: Materials may be edited for content and length.