RichRelevance Launches ‘Deep Recommendations’: The Next Generation of Advanced Commerce Personalization

0 394

RichRelevance, a leader in experience personalization, announced the launch of first-of-its-kind ‘Deep Recommendations’, a set of advanced personalization technologies that, unlike traditional recommender engines, does not need historical events and behavioral data to immediately generate relevant product recommendations.

The new approach solves two problems: (a) it removes constraints associated with traditional recommendations which don’t work for retailers and brands with sparse data – seasonal products, fast-changing catalogs and long-tail products, and (b) it helps product discovery by catching user’s preferences through a product’s visual features and textual description.

Marketing Technology News: Genesys Names New CFO to Drive Next Phase of Rapid Growth

With Deep Recommendations, retailers, and brands that regularly introduce new products can expose shoppers to these new products instantly. In addition, categories such as fashion and home furnishings where shoppers look for ‘visually similar’ or ‘visually complementary’ products can break through the clutter with highly relevant and high conversion visual AI-based recommendations.

RichRelevance Deep Recommendations are enabled by Xen AI, the most advanced machine learning engine in the space and the only one with composite deep learning, an industry-first approach that blends all known data and decisions to predict the next best experience.

Xen AI extracts and combines feature vectors (the “DNA”) found in product text descriptions and catalog images, behavioral data, derived affinities, and stated preferences and matches in real-time with shopper intent to create highly relevant high-conversion recommendations. This helps your customers not only get what they are initially looking for but also inspires them to discover contextual recommendations to fulfill their needs across their shopping journey.

Experience Optimizer (XO), the patented decisioning layer of Xen AI, is used to continuously experiment in order to predict the most favorable outcomes by mixing and matching traditional strategies, personalized strategies, and now, deep learning strategies.

Results from its over 30 early adopters and customers have revealed spectacular results, with Xen AI Deep Recommendations creating an average lift of 40% in engagement and 80% higher attributable sales, in comparison to standard recommendations prevalent in the industry today.

Marketing Technology News: Sitecore Names Steve Tzikakis as Chief Executive Officer

“We instinctively knew that using visual aspects of a product for recommendations is effective in fashion and lifestyle business – it’s much closer to the expertise of our merchandisers. I am excited with early results – our engagement is up 40% over our merchandising rules, and revenue per 1000 impressions has increased by 19%, compared to the other recommendation models,” said Sylvain Lys, Head of Omnichannel Customer Experience at Promod, France.

“Deep recommendations is our top-performing strategy right now and is delivering average attributable sales of Eur 10.68 per click. The results are scarily good. Without RichRelevance, these innovative AI technologies wouldn’t have differentiated us, and helped us grow,” said Anton Paasi, Head of Ecommerce, Verkkokauppa.com, a leading Finnish online retailer.

“Deep Recommendations replicate how store assistants help a shopper with their purchases, by interpreting their likes through a combination of language cues and visual attributes revealed in the shopping journey, along with an understanding of their past affinities to a brand or price point. The relevancy will continuously improve as deep learning algorithms gather more volumes, and Xen AI learns from how users interact with these recommendations,” said Mark Buckallew,  VP, Product Management at RichRelevance.

Marketing Technology News: Merkle B2B Unveiled: Dentsu Aegis Network’s New B2B Powerhouse

Leave A Reply

Your email address will not be published.