Final Report Implementation Online Recommendation Engine
As planned, we successfully completed the “Implementation Online Recommendation Engine” project on December 20, 2018.
What is a Recommendation Engine anyway?
The aim of a Recommendation Engine (automated product recommendation engine) is to motivate customers to (additional) purchases by suggesting the most suitable additional products. Buyers in online shops behave much like customers in retail stores. If you are interested in a product and have selected it for purchase (this happens in the Web shop on our product detail page, PDP for short), you are open to recommendations that might fit the selected product. This increases the probability of a purchase.
What was the content and aim of the project?
The project included an update of the FACT-Finder and an upgrade of the FACT-Finder with new functionalities. This makes it possible to suggest suitable, personalized product and category recommendations to the customer in real time. Based on his search and click behaviour.
This increases our conversion rate (conversion rate = part of the prospective buyers who visit our web shop and become buyers) and the turnover in the shop. We thus profit from automated upselling and cross selling (i.e. from the additional sale of higher-value products or supplementary articles to the articles initially searched for by our customers in the shop or placed in the shopping basket).
Initially, the Recommendation Engine was implemented on the product detail pages for KAISER+KRAFT, gaerner and Kwesto. This corresponds to what you know from Amazon if you have selected a product there. Then you will see “Customers who bought X also bought Y”. We also use this principle.
After that we did the FACT-Finder update. This expansion stage of our Recommendation Engine is based on a self-learning technology (artificial intelligence), in which the individual purchasing behaviour of our customers is continuously analysed. Within the framework of the project, the basis for the further use of this technology was created and an initial basic configuration was carried out. In addition, the various improvements to the FACT-Finder software allow us to optimize search results for our clients. The further “fine tuning” will take place in the course of 2019.
Thorsten Louis sums up the project as follows, “With the integration of the Recommendation Engine, we are taking the digital customer journey to the next level. It enables us to use artificial intelligence (AI) to offer tailor-made product recommendations to the customer at the right time.”