Personalization without interrogation: Shop easier, faster and better, without telling your life story
(PhysOrg.com) -- Whether people are buying headache medication or cars, technology needs to help buyers wade through a staggering number of choices, says a new University of Alberta study.
"Consumers are overwhelmed by choice. We want to make good decisions, and machines and software should be able to help us with that and do it better than they currently do," said Kyle Murray.
Murray and co-author Gerald Häubl, both marketing professors from the School of Business, wrote "Personalization without Interrogation: Towards more Effective Interactions between Consumers and Feature-Based Recommendation Agents," recently published in the Journal of Interactive Marketing. In the study, they examined various studies of shopping patterns, including software with recommendation systems or "agents."
Murray and Häubl found creators of such systems have focused primarily on enhancing the usefulness of these tools to consumers, but devoted insufficient effort to making the agents easier to use.
The authors found a number of existing personalization technologies that provide reasonably good recommendations without "interrogating" consumers. For instance, Amazon.com uses an ever-evolving algorithm that incorporates past purchases with the preferences of similar consumers in a method called collaborative filtering, as well as direct input from users, as a basis for providing customers with individualized recommendations.
But for every good system out there, numerous others exist that aren't very useful, said Murray.
"Technology could screen through all the cars that exist, go through what you want for features and price, look at what's available and then give you a handful to look at," he said.
"There are basic things that could be done to make them better, more human," said Murray. For example, software recommendation systems have historically asked far more questions than a human advisor would before making a recommendation.
Since humans tend to treat computers like people, the authors of the study state that it is reasonable to suspect that computers-and, in particular, recommendation systems-would be more attractive to consumers if they acted like people.
"People like systems that incorporate human gestures and appearances and interact with users in a polite and empathetic manner. Technologies such as natural language and ambient intelligence may also be important in this regard," said Murray.
Murray and Häubl also looked at research that showed how having too much choice can lead to other negative consequences, such as increased regret, decreased product and life satisfaction, lower self-esteem, and less self-control. Related research has shown that, although consumers may prefer to buy products with more features and capabilities, ultimate satisfaction with their purchases decreases to the extent that these very features make products more difficult to use.
"These findings are not just important for buyers; this type of decrease in consumer satisfaction tends to also have a negative impact on sellers' long-term profitability," said Murray.