Semantic technology increases value of online reviews

Share this article:

For the average consumer, online product research is a double-edged sword. While Web 2.0 platforms have facilitated the creation and distribution of user-gener­ated reviews, sorting through these articles and comments can be a daunting task for the consumer. The very same technol­ogy built to improve the Web has actually made the Web a bit more unwieldy.

Fortunately, a new wave of Internet start-ups intends to make life easier for the online consumer. Pluribo, an early stage semantic technology venture, has recently developed a Firefox plug-in that scans multiple user reviews on a site such as Amazon.com to produce a “boil-down,” or a condensed review based on the aggre­gate. At the core of the firm's belief is that automation can, and should, humanize the Web.

“The promise of automation should help you understand the why — that is, why something is good or bad,” said Samidh Chakrabarti, co-founder of Pluribo. “It is empowering.”

And while blog posts, discussion forums and comments empower customers to put in their two cents, the sheer mass of content can be very discouraging for cus­tomers. Furthermore, many customers are concerned with the authenticity of some reviews. Chakrabarti noted that Pluribo's technology is quite good at detecting fake reviews, and is pleased to take this com­ponent out of the equation.

“The argument that marketers can manipulate the conversation is a anti­quated view,” he said. “The role of the marketer should be to listen.”

Sam Decker, CMO of Bazaarvoice agrees. His firm's user review technol­ogy also yields vast amounts of mean­ingful insight. Decker noted that 85% of reviews are positive, contrary to the popu­lar belief that only unhappy customers write reviews. Also, most customers write reviews in the interest of helping other customers, as well as the companies.

“What we have learned over the past two and a half years is how organizations adopt technology and how customers tick,” Decker said.

Semantic analysis of these reviews can then help the savvy marketer amplify busi­ness decisions. “A retailer is an editor of products,” Decker added.

He suggests using semantic data to inform which products are kept in stock or which features resonate best with custom­ers. “Companies such as Home Depot, Dell and Macy's realize that marketing can't hide product flaws,” said Decker, commenting on some of his firm's 220 clients.

With semantic analysis, online reviews might appear to be closer to a customer service function than to marketing. Yet for the time being, Decker said that his firm's services tend to be within the mar­keting spend.

“Marketers want reviews because they know it has ROI,” Decker said. “And then someone gets their hands on the data and realizes that it is their best friend.”

This material may not be published, broadcast, rewritten or redistributed in any form without prior authorization. Your use of this website constitutes acceptance of Haymarket Media's Privacy Policy and Terms & Conditions