Is 90% Good Enough?
VIBE Alpha Test
The first Vibe® Engine, presented at the 2005 Shop America Conference, was tested on a pool of aloha shirts (12 men’s and 12 Women’s) provided by Hilo Hattie of Honolulu Hawaii. 80 conference attendees were invited to take an interactive version of Vibe® Personality Test that generated product recommendations in the form of personalized discount coupons. Next, before shoppers were given coupons, they were asked to select a couple of preferred shirts from a poster that depicted the pool of shirts. Results showed that Vibe® recommendations were approximately 90% similar to shopper preferences.
Compare & Contrast this result with other recommendation approaches that use only buying history and key words to second-guess what one might be interested in. For example, at some book sites, if you purchased a book on “kite flying” for a friend, you’d likely get recommendations for kite flying books long time into the future – which could be pretty annoying and certainly not all about you.
But the Truth Is that even product recommendations merely based on demographics such as gender, are currently a big win on the Internet. For example, imagine clicking a woman clicking a product and all new features are re-generating as women’s products… Not bad, but online retailing should aspire to more which is why Vibe® Technology exists today.
For more, see Technology, How It Works, Theory, Efficacy, Benefits, FAQs.