Insights from the Data Mining Challenge:
Targeted voucher allocation in e-commerce

27 Jan 2026

Ramon Schleicher is studying business informatics and delivered the best result in the Data Mining Challenge 2025

As part of the Data Mining Challenge, students were once again able to delve deep into a data set this semester. The aim was to send vouchers to customers of an online shop in such a way that, if possible, only those customers who would otherwise be unlikely to return would receive a voucher.

 

The challenge of customer loyalty in the digital age

In today's digital world, customer loyalty plays a crucial role in the success of online shops. The cost of online advertising is often high, and revenue from initial purchases usually contributes little to a company's bottom line. This is why enterprises are looking for effective strategies to retain customers in the long term. One promising solution is the targeted awarding of vouchers - an approach that took centre stage in the Data Mining Challenge.

 

The Data Mining Challenge: innovation in the voucher strategy

In the current semester, students took on the challenge of sending vouchers in such a way that they only reach customers who would otherwise not return. The idea of issuing a voucher to every new customer was challenged by using intelligent forecasting models.

 

Forecasting models for returning customers

As part of the challenge, the students developed various data mining models. These helped to gain valuable insights from real data from an e-commerce enterprise. A central aspect was the identification of customers who would not return without an incentive. Here, new variables derived from the data set proved to be particularly important in order to feed the models with specific knowledge.

 

Success story: Mr Ramon Schleicher and his data model

This year's winner of the Data Mining Challenge is Mr Ramon Schleicher, a student of business informatics at the department of applied computer sciences at Fulda University of Applied Sciences. With his model, he was able to predict non-returners more accurately than all other participants. One of his key findings was that customers who order multiple products tend to return more often. Through the skilful selection and combination of variables, Mr Schleicher was able to develop a model that went beyond the usual basis for decision-making and achieved excellent results.

 

Conclusion: The future of customer loyalty with data mining

The results of the Data Mining Challenge show that significant improvements in customer loyalty can be achieved through sound data analysis and creative approaches. Interested enterprises and students are invited to take part in future challenges and discover new ways of approaching customers.

 

We congratulate Mr Schleicher on winning this year's Data Mining Challenge.

back