When Overstock Art reached out to Infinite Analytics, they were facing a problem. Half of their paintings are based on renowned artists like Monet, Renoir, and Klimt. The other half of their art was from lesser known contemporary artists. The relatively unknown artists produce beautiful work, but Overstock Art didn’t have an efficient way to connect customers to relevant art when they didn’t specifically search for a piece.
The company knew it could improve its business if it could find a way for customers to discover the perfect painting in as few clicks as possible.
This was when Infinite Analytics, a predictive marketing company, got involved. To effectively personalize, you need data, and in this case, little was initially understood about the paintings from contemporary artists. Infinite Analytics employed a variety of technologies for product catalog enrichment to improve personalization, even when the product was new or had little information in the product description.
Here are examples of some of the technologies used for overstockart.com to provide personalized recommendations:
Natural Language Processing: Product descriptions are processed to understand the meaning and structure of both structured and unstructured data. This allows for increased relevance and is a foundation for other kinds of semantic enrichment.
Semantic Technologies, Knowledge Base Enrichment: Data from the web, including relevant articles, encyclopedic references, or other websites, are used to understand more about an artist and the style or genre for the painting. Social profiles about the artist are crawled and used to further enrich the understanding of the artist and their style.
Image Recognition: All the product catalog photos are processed to understand the exact breakdown of colors and patterns. This is used to prioritize compatible colors and understand related works. This has many applications when merchandising apparel, but this is even clearer in this case when the products are so beautiful and full of colors.
Selling only a portion of the product catalog is a common problem. Most e-commerce retailers sell a fraction of what is available to purchase on their website. Legacy product recommendation companies have done little to address this issue because they rely largely on past behavior to drive personalization. This has led to spikes in demand for certain products, but overall has failed to deliver on the promise of personalization for modern e-commerce companies. Without data, many companies are stuck manually promoting products or using hard coded recommendations. This can be a time consuming process, or when your large product catalog has a lot of turnover, it can be impossible to keep up, which results in lower conversions. Infinite Analytics provides product catalog enrichment as a service that counteracts this issue, and drives incredible results for our customers.
Infinite Analytics creates personalized experiences by using all available data about the customer, enriching the product catalog, and incorporating context and data from the web. This multi-sided approach allows for effective personalization even when there is initially little information. A recent A/B test for a recent customer showed this to be incredibly effective as the averager number of clicks until purchase was lowered from 120 to 80. The effect is a dramatic increase in conversions and engagement.
Find out how we can help you create an extraordinary experience for your customers by emailing us at firstname.lastname@example.org.