Blue Tomato

  • RETAIL SEGMENT: Sporting Goods
  • PRODUCT: Personalized Recommendations
  • CHALLENGE: Wide range of products. 60% Webshop in 14
    languages and nine currencies
  • RESULTS: 3x revenue from orders containing product recommendations. 20% increase in
    average shopping basket. Average of one more product per purchase. Improved revenues on mobile devices

Blue Tomato was founded in 1988 as a snowboard school by the former European Snowboard Champion Gerfried Schuller. Since then it has transformed into a successful international boardsport and fashion shop. Blue Tomato now owns more than 30 shops in Austria, Germany and Switzerland and their webshop offers more than 450,000 products from more than 500 brands and delivers to more than 40 countries. Having launched its webshop in 1997, Blue Tomato was an eCommerce pioneer. It now aims to become the leading omnichannel retailer for boardsports and freeski in Europe.

Product variety proved too challenging for existing recommendation engine

By 2016 the growing breadth of Blue Tomato’s product range became a challenge for the company’s existing recommendation engine.

“Our traditional tools weren’t working for us anymore,” explained Andreas Augustin, Head of Webshop Development at Blue Tomato. “Our existing tool had reached its capacity to make automatic recommendations. We spent a lot of time and resources trying to manually improve results and ingest products, with limited results.”

Blue Tomato therefore sought a more sophisticated recommendation engine to handle its growing complexity. “We knew of several diverse eCommerce tools that included an element of machine learning within its platform to make recommendations”, said Andreas. “We decided we did not want this type of solution, but one whose core competency was recommendations.”
During an exploratory stage Blue Tomato evaluated solutions from six different vendors and finally chose Recommend™ by RichRelevance for its advanced machine learning algorithms, ease of use and merchandising functions.

Machine learning algorithms that challenge each other

“We particularly liked RichRelevance’s ‘King of the Hill’ approach, with its different machine learning algorithms that continually challenge one another to get the best results,” said Andreas Augustin. “We also valued the extended merchandising functions that enable us to maintain specialist product areas inhouse. Finally, the personalization developed specifically for mobile devices was very important for our omnichannel strategy.”

Experienced RichRelevance consultants supported Blue Tomato through the implementation process, which ran smoothly and fast despite the webshop’s complexity. “Thanks to RichRelevance’s great support – often on short notice – we were not only able to optimize the system but could also be assured that our personalization project would be successful in real-time mode,” explained Andreas Augustin.

Personalization that customers and staff can equally identify with

Blue Tomato’s revenue created by product recommendations has grown significantly since going live with RichRelevance in Autumn 2016. Crucially for Blue Tomato, revenue resulting from product recommendations has tripled, proving the value of the investment in RichRelevance. That the personalized recommendations resonate with Blue Tomato’s customers is also rejected by the fact that they spend more. “Thanks to RichRelevance, the value of the shopping baskets resulting from the product recommendation has increased by an average of 20 percent, with an average of one more product purchased by each customer”, summarized Andreas Augustin. ”The numbers apply as well for the recommendations shown on the mobile devices, where significantly less products can be listed but thanks to RichRelevance these are the most relevant.”

Blue Tomato’s product management and marketing teams are also impressed by the the quality of the recommendations. One type of recommendation problem that has been historically difficult to manage is when customers buy separate matching products – for example bikini tops and bottoms in different sizes or from different collections. When this happens, appropriate recommendations should still appear as if the customer had chosen a matching set. “Before using RichRelevance, this common scenario had been difficult to maintain and manage,” explained Andreas Augustin. “We were able to configure Recommend™ to ensure matching parts were listed together. However, this wasn’t even necessary as the algorithms figured it out pretty fast themselves.”

More personalization, also for content

Looking to the future Blue Tomato plans to extend personalized recommendations to its various online theme parks such as “beach life,” and also for the Blue Tomato “rider crew” sites that feature snowboard, freeski, surf and skate athletes who are sponsored by 30-40 different brands. The company plans to use RichRelevance to personalize the content that visitors see, along with more interactive sites to further improve the customer experience.

Share :