August 06, 2019
Want to Boost Sales? Make Sure Those Five-Star Reviews Focus on Price, Aesthetics, Research Shows
New research from Carnegie Mellon’s Tepper School of Business suggests that showcasing these factors in online reviews has the same impact on sales as a price cut.
When customers are scanning product reviews to decide whether to make a purchase, the dimensions that matter most to them are price and aesthetics, according to new research by the Tepper School of Business.
An analysis of more than 500,000 reviews across 600 different product categories and 250,000 online users yielded the results for the research from Dokyun Lee, Assistant Professor of Business Analytics, and his coauthors — Kannan Srinivasan, H.J. Heinz Professor of Management, Marketing, and Business Technologies and Xiao Liu of New York University’s Stern School of Business.
Published in the Journal of Marketing Research, the paper is titled “Large-Scale Cross-Category Analysis of Consumer Review Content on Sales Conversion Leveraging Deep Learning.” The paper studied data that tracks product search, customers’ review reading behavior, and conversion — typically a closely guarded resource among online retailers — from one of the largest review platforms available.
Lee and his coauthors used deep learning text-mining techniques to study several dimensions, including brand, feature, conformance, and durability. Ultimately, they found that no matter what the product was, price and aesthetics held the most sway over whether a person bought the product.
Most consumers read fewer than 10 reviews before making a decision, so it doesn’t take much for a positive mention of these factors to tip the scales. Simply re-ordering them so the ones mentioning price and aesthetics are on top can drive an increase in sales, at least in the short term.
“If you swap out one review that person reads with a review that has positive comments on price or aesthetics, but keep everything else the same, it has the same positive effect on sales as up to a 1.6 percent price cut,” says Lee.
However, consumers also want some diversity in the content they read — meaning they prefer reviews to contain information about other factors, they just want to read about price and aesthetics first. Sales will be boosted in 66 percent of cases where reviews are optimized to reflect these preferences, Lee said.
The research is the first to actually dive into the content of consumer reviews when connected to clickstream and conversion data, thus illustrating how content influences the decision to purchase.
The impact of reviews on sales was highest when the product’s average rating is higher and there is little variance in the ratings, Lee says. In other words, when there is a consensus among customers that the product is good or review rating inflation, review content matter more.
Consumers rely more on reviews when a product’s market is more competitive, is immature, or when brand information is not easily accessible, the paper also finds.
The research showcases how to use deep learning techniques to extract insights from unstructured text data for business insights. Lee believes that there is a big underutilization of unstructured data such as text, images, and videos and much economic values to be extracted.