Arvato CRM Solutions collaborates with University of Münster on real-life retail customer segmentation
Guetersloh / Münster – a project team of two research fellows from the ERCIS Omni-channel lab at the University of Münster has applied a stream clustering to tackle customer segmentation challenges in omni-channel customer relationship management. The two researchers, with a focus on big data architecture and data analytics, developed an approach to cluster customer segments by using an advanced analytics algorithm within just three months. The project is based on a real-life case of a Spanish retailer in the home furnishings and textiles sector.
Clustering is a statistical methodology that identifies homogeneous groups of objects and allows the discovery of previously unknown patterns within data. Clustering methodology has been applied in multiple customer service scenarios, such as customer segmentation, social media monitoring and fraud detection.
A key challenge for customer segmentation is that it is usually just a snapshot analysis, that identifies segments at a specific point in time. This ignores the fact that customer segments are highly volatile and segments change over time as trends come and go and preferences adapt. Once a segment has changed, marketing strategies can be made unusable.
To alleviate this problem, our project team has developed an innovative segmentation approach based on stream clustering. In this case, the approach allows us to identify and monitor segments over time – for example, based on a stream of online transactions. The algorithm then incrementally updates segments so that they can be identified in real-time. Therefore, the results can be used at any point without the need for complicated recalculations.
In total, more than 1.7 Million transactions have been analysed and customers have been categorized into multiple clusters, such as loyal customers, online returns, and large orders. This is based on distinguishing characteristics such as online ratio, order size, frequency, and return rates etc. By applying this stream clustering methodology, the analysis process is faster, easier to apply and also allows to track how segments have changed.
Jiaqing Zhong, senior manager strategy execution, Arvato CRM Solutions said, “Our clients really value Arvato’s deep vertical market expertise. In this specific case, we have come up with solutions to better segment specific targeting especially focused on online behavior - an area that’s of special interest to our clients. It is great to collaborate with ERCIS on this project in developing a state-of-the art algorithm. We are very much looking forward to integrating this prototype into our analytics solution offering.”
The ERCIS Omni-Channel Lab, at the University of Münster, is a partnership between ERCIS and Arvato CRM Solutions. The Lab combines the academic research and teaching activities of ERCIS and Arvato’s practical experience of delivering tech-enabled Omni-Channel CRM solutions from 110 global locations for many of the world’s best-known brands. Its areas of investigation focus around ‘Processes’, ‘Data’ and ‘Analytics’.