The story started when Coop needed to carry out specific, far-reaching analysis work in order to be able to focus and refine its offers to its customers. Enfo agreed with them that advanced processing using R on SQL Server 2016 would be the best option for the analysis work. In addition to an entirely new arrangement of SQL Server 2016 with R services, Enfo established a full integration and process operation in which the data is collected and handled from its original data warehouse to the analysis environment. Coop was also helped to configure a model in R to analyse, for example, which parameters are important when a specific offer is taken up. The parameters are matched against the existing customer profile and the result gives the probability that an offer will be taken up in respect of each individual customer.
"We did not just want to implement a model, we wanted to create a platform on which we can carry out advanced analysis work on a large scale" - Fredrik Khan, Enfo
The result of Enfo's project with Coop is a fully integrated R environment where the customer today reviews a majority of their offers which are then sent out to their customers. All the information regarding the take-up of the offers is used to further train the model, in other words machine learning. The result is that the offers that appear on the internet, on mobile phones and in the membership magazine “Mersmak” now have a much higher probability of being used by customers! In addition Coop now have a secure and future-proof platform to run even more models and processes on.
"It was surprisingly simple to integrate the platform into an existing Microsoft environment; we completed the project in a fairly short time and under budget despite its scope being widened" - Erik Perninge, Enfo