Researcher of the Month
Online feedback: How the wrong system can distort the picture
Especially online shoppers often base their purchasing decisions on customer reviews. Hardly anyone books a hotel without first scrolling through all the feedback posted by previous guests. In his research, WU Professor Ben Greiner, head of the Institute for Markets and Strategy and the Competence Center for Experimental Research at WU, investigates the factors that cause distortions in online feedback systems. Such distortions are detrimental not only to customers, who are provided with less reliable information, but also to businesses, because their customers are less likely to trust them due to less reliable information.
Whether it’s booking a hotel or shopping online: Many people base their decisions on customer reviews. At the same time, publicly posted reviews provide a strong incentive for companies to improve products and services. For this to work, customer feedback has to be both honest and informative. WU Professor Ben Greiner uses field data analysis of online marketplaces like eBay and laboratory tests to investigate which online feedback systems work best. He observed a number of different patterns of social behavior that were affected by the feedback rules of each system.
Revenge feedback for negative reviews
On eBay, one of the market platforms that allows buyers and seller to provide mutual feedback, Greiner’s study provides evidence for strong reciprocity in feedback behavior. Positive reviews from buyers were almost always rewarded with positive feedback from vendors, whereas a negative review almost always resulted in negative revenge feedback from the seller. As a result, many people who had a negative experience with a purchase chose not to leave feedback at all, for fear of getting negative feedback in return. This distorts the entire feedback image: 98% of the feedback given on eBay is positive. “This means that the collective feedback offers only a limited amount of information,” says Greiner, “Our laboratory studies have shown that this is the result of the system’s openness. If you change the rules and don’t publish feedback until both parties to the transaction have submitted their review, it leads to more negative, honest feedback but, unfortunately, also to a lower rate of participation in the feedback system as a whole. It also works better if only one-sided feedback is allowed.”
Allowing posters the option to retract feedback may also have unintended effects. Ideally, this option allows both parties to the transaction to find a solution in the event of a conflict. The studies revealed the opposite behavior, however: For example, sellers who were expecting negative reviews posted negative feedback about buyers as leverage to later convince the buyers to retract their own negative review. “This also results in feedback distortion. The information value and effectiveness of the feedback system are considerably reduced,” says Greiner. Overall it was shown that the specific rules of individual feedback systems determine the strategic incentives buyers and sellers are exposed to when they contribute their feedback. An efficiently designed and planned feedback system takes these factors into account.
For his research, Ben Greiner used field data from eBay and other online market platforms as well as data from laboratory experiments. Greiner explains, “In the field, there are a number of variables we can’t observe. For this reason, we recreate markets in the computer laboratory and ask human participants to make decisions taking the roles of buyers and sellers. The profits that participants made in the lab experiments were paid out to them in cash. This ensures the authenticity of the decisions made. Like in a wind tunnel, we can then change the design, i.e. the rules of the feedback system, and observe the reactions of the buyers and sellers.”