Relevante website en media met algoritme met Raptor

Bog & idé – More Inside – Case Study

Recommendations that convert 

In order to optimize user experience and customer loyalty, Raptor is delivering personalized product recommendations, which are engaging, converting, and easy to scale across an entire platform. 

Bog & idé – More Inside – Case Study

Reading time: 5 minutes

Bog & idé is the largest retailer of books in Denmark. Their value proposition is based on the needs and wants of their customers, and to support those values they are using Raptor Smart Advisor to recommend the most relevant products to their online customers.

Furthermore, Bog & idé is developing a tool, which will allow in-store personnel to utilize real-time data from Raptor Services. This feature will reveal relevant up- and cross-selling opportunities in relation to all “Click & Collect” customers in physical stores. 

Currently, 65% of all online orders are collected in-store, so there is an enormous potential in personalizing an omnichannel-solution based on data from Raptor.

Established in 1969 as Bogpa with 36 booksellers, Bog & idé is now operated by Indeks Retail A/S with 125 stores. 

A growing online presence represents a very important source of income for Bog & idé. Therefore, they always strive to meet the different needs of different customers, both online and in-store. As a result of their efforts, Bog & idé was recently awarded the 3rd place for “Best Omni-channel E-Commerce business” at FDIH’s “E-Commerce Awards.”

Recommendations that convert 

In order to optimize user experience and customer loyalty, Raptor is delivering personalized product recommendations, which are engaging, converting, and easy to scale across an entire platform. 

To ensure that every online interaction is as personal as possible, every recommendation from Raptor is calculated in real-time, which means that the customer’s own decisions are helping to shape their individual customer journey. 

Revenue lift increase from 3.74% to 15.83% in three weeks

Following the initial integration, we gave the algorithm some time to learn from clickstream data, and after the first week, revenue lift had increased from an initial 3.74% to 9.79%. 

In close collaboration with the team at Indeks Retail, we started a process to fine-tune and customize the solution on Bog & idé’s online platform. 

By doing so, we achieved a more advanced synergy between our recommendations and the customer journey on their site. One week later the revenue lift had increased to 15.83% with Direct Revenue at 5.66%. 

Revenue by Raptor explained

Revenue by Raptor is defined as the revenue generated directly and indirectly from Raptor Modules. 

Income from customers that click on and buy a product from a Raptor module registers as Direct Revenue, whereas income from customers that click a product from a Raptor module and then purchase something else, register as Assisted Revenue. Together they make up Revenue by Raptor. 

Timeline and results

At this point, Raptor Smart Advisor is recently implemented. The modules are on the website, but the algorithm has not yet mined large amounts of data. Therefore, the patterns calculated by the algorithm are less intricated, and the recommendations are less accurate in the first few days after launch. 

A week later the algorithm has mined clickstream data and mapped the co-relation between products and user behavior. The modules on the site now recommend products that are much more relevant to the individual user than before. 

Just one week after the optimization process, the total Revenue by Raptor had increased to 15.83% of total online revenue, compared to 3.74% Revenue by Raptor two weeks before the optimization. After the optimization process, Direct Revenue increased more than double from 2.37% to 5.66%. 

Direct Revenue is a major factor in terms of how the modules are performing on the site. The more products are sold directly from recommendations, the more effective the solution is. The increase in Direct Revenue is also a clear indication that the recommendations are more accurately tailored to meet the needs of the individual user. 

Key takeaways

• As our algorithm continuously learns from behavioral patterns, the performance increases over time. This is because our recommendations become more qualitative and relevant for the individual end-user.

• When we collaborate closely with our customers we understand their individual business needs, and that is the key for both parties to succeed. Therefore, working closely with our customers is what makes the difference, when it comes to performance and Return on Investment. 

• Volume of data is essential for any machine learning process. The more data we have, the better the recommendations will become.
  
  Voor meer informatie, Tjerk van der Veen, tjerk@ecommerceweb.nl, 0638882222, reseller Raptor persuasion software