A New Algorithm is about to Mine Your Product Reviews, then Upsell You. Call it Computational Advertising, and it Works on Amazon & Twitter
By Lori Cameron and Michael Martinez
Published 09/08/2017
Share this on:
Researchers have devised a new algorithm that now exploits—at a fine-grained level—whatever you write on a product review site and then mines those opinions in an effort to target you with better, more relevant advertising pitches.
It’s another way that advertisers and marketers are trying to connect with consumers. It’s also a new field that merges computing, economics, and business in a revolutionary way. Call it customized advertising.
Mauro Dragoni, a researcher scientist at Fondazione Bruno Kessler, has already tested his new model on Amazon product reviews and leveraged his results to create advertisements for Twitter.
His team saw a 12% jump in user engagement when the messages hit Twitter timelines.
“The approach has been validated in a real-world scenario, with the results demonstrating how the analysis of user-generated content provided by a specific community can be exploited to build attractive messages,” Dragoni writes in his study entitled “A Three-Phase Approach for Exploiting Opinion Mining in Computational Advertising.”
The workflow of the presented three-phase approach: (a) extracting aspects from user-generated content and creating clusters with semantically correlated labels, (b) computing aspect polarities, and (c) managing users profiles to infer the most interesting aspects to be exploited. User-generated reviews are used as input for the first two phases. The third phase combines the outcomes of the first two with the user profiles built from Twitter timelines. The merged data serves as input for the message generator component that produces the final set of tweets created to attract users.
“The novelty this work brings is linked to the exploitation of users’ perspectives to improve their engagement when visiting product pages,” he adds.
What sets his proposed model apart is how it looks at why you, the consumer, prefer a particular product. This opinion-mining goes beyond the old algorithms that typically don’t do much more than mine social media and product reviews to find out what people like, track the geographical location of customers, and monitor their purchase histories to improve advertising campaigns.
Those previous algorithms didn’t dive deep into why people purchase some products and not others. They do not discover the specific features people love most about the products they buy.
Therein lies the key.
If businesses know exactly what consumers love about their products, they can turn those features into hot selling points in future advertising campaigns.
“Outcomes reported here demonstrate the viability of the proposed solution and open interesting directions for future research, in particular, the possibility of applying such techniques in advertising campaigns, where knowledge extracted from user generated content can be exploited to better focus campaign component design,” Dragoni writes.
Related research from the Computer Society Digital Library
Lori Cameron is a Senior Writer for the IEEE Computer Society and currently writes regular features for Computer magazine, Computing Edge, and the Computing Now and Magazine Roundup websites. Contact her at l.cameron@computer.org. Follow her on LinkedIn.
About Michael Martinez
Michael Martinez, the editor of the Computer Society’s Computer.Org website and its social media, has covered technology as well as global events while on the staff at CNN, Tribune Co. (based at the Los Angeles Times), and the Washington Post. He welcomes email feedback, and you can also follow him on LinkedIn.
