Thesis project: Product Recommendations based on Customer Segmentation and Market Basket Analysis in an online environment
2014-02-27
Our company develops an online tracking system used by e-commerce companies in Scandinavia. The system facilitates decision makers to optimize ROI in regard to marketing banners and messages, which results e-businesses making advantageous decisions and increase their overall profit margins generated by online marketing campaigns.
An additional feature to gain value from the tracking system is to identify and show relevant product recommendations to visitors based on customer segments, market basket analysis, and other customer and historically data.
A thesis project which describes and develops a framework that creates customer segmentation and uses MBA in a dynamic real-time web based environment would be an interesting and valuable project for our company as well as for you as a student.
We are able to provide large data sets containing data related to the visitors – for instance timestamps, geographically information, devices, referring websites, search keywords, languages, navigation on websites, viewed products and categories, sub goals such as “add to card”, conversion rates etc.
The aim of the project is a model generating product recommendations, which increases the profit of each visitor. Speed is an important factor and since the tracking system is using several caching methods the model cannot be executed for each visitor. One way to handle these speed challenges is to classify the visitor based on predefined customer segmentation and find and rank predefined product recommendations based on the segment classification. Since manual work is a potential risk for failure, the model should either be based on pseudo code that can be translated into PHP (a web programming language) or some kind of open source command line software that runs on Linux and outputs structured data based on input data in a batch job.
If you find our project interesting, we would appreciate to hear from you. Please send Anders Lindgren an email with relevant academic information about you, and a short description telling us why you think you are the right candidate for the project. If we find a match, we will invite you to a meeting in which we can discuss the project further.
Any questions can be directed to Anders Lindgren at anl@reaktion.dk.
Best regards,
Reaktion ApS
Anders Lindgren
Co-Founder, MSc Logistics & SCM @ BSS