• Omni-Channel CRM

    Omni-Channel Processes
    Omni-Channel customer service demands the orchestration of all communication channels, so that customers can use the one(s) they prefer, and the company has a comprehensive view of the customer. This necessitates processes that enable channel alignment and customer identification regardless of the specific platforms or systems used. Process reference modelling is a means to tackle this challenge and a cornerstone of the Lab’s work.

    Data Integration
    Data integration enables heterogeneous data sources to be combined into a unified view to get new insights and improve customer experience and internal processes. The complexity of data integration in an Omni-Channel CRM context is driven by the fact that clients use different solutions. Moreover, as clients add new communication channels the number of data sources increases, and the representation of the data can. The Lab investigates big data technologies to propose database management architectures to tackle these issues.

    Analyzing customer and operational data across all channels has become a key driver of business success. A particular interest lies in the identification and analysis of market segments, i.e. the separation of a heterogeneous market with diverse preferences into subsets of homogeneous groups that share similar interests. The identification of customer segments makes it possible to target each group with appropriate value propositions and marketing strategies.

    • Carnein M, Heuchert M, Homann L, Trautmann H, Vossen G, Becker J and Kraume K (2017), ‘Towards Efficient and Informative Omni-Channel Customer Relationship Management’. In: Proceedings of the 36th International Conference on Conceptual Modeling (ER'17). Valencia, Spain, 2017 pp. 69-78.
    • Heuchert, M., Barann, B., Cordes, A.-K., & Becker, J. (2018). An IS Perspective on Omni-Channel Management along the Customer Journey: Development of an Entity-Relationship-Model and a Linkage Concept. In Proceedings of the Multikonferenz Wirtschaftsinformatik 2018, Lüneburg, Deutschland.
    • Heidekrüger, R., Heuchert, M., Clever, N., & Becker, J. (2018). Towards an Omni-Channel Framework for SME Sales and Service in the B2B Telecommunications Industry. In Proceedings of the Multikonferenz Wirtschaftsinformatik (MKWI 2018), Lüneburg, 386–397.
    • Trautmann, H., Vossen, G., Homann, L., Carnein, M., & Kraume, K. (2017). Challenges of Data Management and Analytics in Omni-Channel CRM. In Becker, J., Backhaus, K., Dugas, M., Hellingrath, B., Hoeren, T., Klein, S., Kuchen, H., Trautmann, H., & Vossen, G. (Eds.), ERCIS Working Papers: Vol. 28. Münster: European Research Center for Information Systems.
  • Customer Segmentation using Stream Clustering

    Customer Segmentation is one of the most important tools in marketing. It aims to identify groups of customers that share similar interest or behaviour. These segments are commonly used in marketing in order to target customer segments with tailored marketing strategies and unique value propositions. Virtually every marketing department uses some form of customer segmentation often based on pure intuition and experience. More mature marketing departments, however, make use of data driven techniques such as cluster analysis in order to identify better segments. This analysis is typically performed as a snapshot analysis where segments are identified at a specific point in time. However, this ignores the fact that customer segments are highly volatile and segments change over time. Once segments change, the entire analysis needs to be repeated and strategies adapted. For this reason, the Omni-Channel lab is researching stream clustering approaches in order to identify and track customer segments over time. Stream clustering is an extension of traditional clustering. These algorithms can can adapt to changes over time and identify emerging segments while forgetting outdated ones.

    The Lab has a longstanding history in this field. Its contributions include extensive information about currently available techniques and algorithms published in a survey paper. Additionally, the performance of algorithms was empirically compared in order to identify strength and weaknesses of existing approaches. Based on this information, we have developed a new stream clustering algorithm which is vastly more efficient in its cluster generation and brings the entire process closer to real time applications. Currently, we are applying these approaches to customer segmentation and evaluating them based on real customer information.

    In addition, the Lab has a lot of expertise in the analysis of textual data. Traditional algorithms to analyse textual data suffer from the same problems as mentioned above. That is, they do not adapt to changes over time and cannot be used to analyse streaming data. For this reason, we have developed new approaches which analyse stream of textual data. Our algorithms allow to extract the currently most relevant topics over. This allows to keep track of currently "trending topics" or common issues , e.g. in social media. Application scenarios can include the improvement of customer service and brand management.

    • Carnein, M., & Trautmann, H. (2019). Customer Segmentation Based on Transactional Data Using Stream Clustering. In Proceedings of the 23rd Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD '19), Macau, China. (Accepted)
    • Carnein M. and Trautmann H. (2018), "Optimizing Data Stream Representation: An Extensive Survey on Stream Clustering Algorithms", Business and Information Systems Engineering (BISE). (Accepted)
    • Carnein M. and Trautmann H. (2018), "evoStream - Evolutionary Stream Clustering Utilizing Idle Times", Big Data Research.
    • Carnein M., Assenmacher D. and Trautmann H. (2017), ‘Stream Clustering of Chat Messages with Applications to Twitch Streams’. In: Proceedings of the 36th International Conference on Conceptual Modeling (ER'17). Valencia, Spain, 2017, pp. 79-88.
    • Carnein, M., Assenmacher, D. and Trautmann, H. ‘An Empirical Comparison of Stream Clustering Algorithms’. In: Proceedings of the ACM International Conference on Computing Frontiers (CF ’17). Siena, Italy, 2017, pp. 361–365.
  • Social Media Analytics

    Social Media provides an abundance of information about people, companies as well as societal and political developments. Analysing and understanding this kind of data is key to provide the best service to customers. For example, the Lab analysed customer service in social media. When customers have a question, request or complaint about a service or product, they often turn to the company for help. To streamline this process, companies usually provide dedicated customer service where employees respond to these inquiries. Traditionally, customer service was provided via phone or mail, however an increasing number of companies expand their service to more modern channels such as Facebook or Twitter. However, the quality of this service varies vastly. While some companies respond within seconds, others take days. The Lab has performed an extensive evaluation of the different social media strategies. In particular, we evaluated hunreds of millions of service requests in social media and analysed how fast they received a response. One finding of the study was that the response times via the social media channels are considerably faster than via traditional communication channels such as email. In addition, there are considerable differences in how fast companies respond. While some answer within minutes, other take days or weeks. 

    The Lab has also carried out various other projects regarding the analysis of social media data. These projects include topic detection from social media data as well as the extraction of customer preference from social media posts.


    • Carnein, M., Homann, L., Trautmann, H., Vossen, G. and Kraume, K. ‘Customer Service in Social Media: An Empirical Study of the Airline Industry’. In: Proceedings of the 17th Conference on Database Systems for Business, Technology, and Web (BTW ’17). Stuttgart, Germany, 2017, pp. 33-40.
    • Carnein M., Assenmacher D. and Trautmann H. (2017), ‘Stream Clustering of Chat Messages with Applications to Twitch Streams’. In: Proceedings of the 36th International Conference on Conceptual Modeling (ER'17). Valencia, Spain, 2017, pp. 79-88.
  • Recommender Systems

    Recommender systems aim to provide personalized suggestions to customers which products to buy or services to consume. Traditionally, recommender systems use the purchase history of a customer, e.g., the purchased quantity or properties of the items. While this allows to build personalized recommendations, it is a very limited view of the problem. Nowadays, extensive information about customers and their personal preferences is available which goes far beyond their purchase behaviour. For example, customers reveal their preferences in social media, by their online search behaviour or their interest in specific newsletters.

    The Lab has carried out multiple real-life cases where it developed innovative methods to improve these recommendations. Our strategies combine information from various different sources in order to identify the customer's preference better. In addition, our approaches make use of social factors and friendship information. The lab is also evaluating different architectures and design choices for recommender systems in order to to improve the performance of recommender systems.


    • Homann, L., Maleszka, B., Martins, D., & Vossen, G. (2018). A Generic Framework for Collaborative Filtering Based on Social Collective Recommendation. In Proceedings of the International Conference on Computational Collective Intelligence (ICCCI 2018), Bristol, UK.
  • A reference model for BPO providers in Omni-Channel CRM

    To fully understand and structure the processes behind Omni-Channel CRM, our goal is to develop a reference model that captures its core management, support and operational delivery activities.

    Our initial investigations suggest that a model built upon a ‘house’ could be a pragmatic approach - management processes are situated in the ‘roof’, the primary part of the building is made up of business processes, and support processes provide the foundation, as they are necessary to run the business. Building on the icebricks methodology, the framework and main process level is specified completely, while the customer-facing processes also cover detail process level.

  • Touchpoint Management and the Customer Journey


    Customer interactions are changing in the digital age with implications on the digital enterprise. The predominant channel-based thinking is superseded by touchpoint-based thinking, which entails a more granular perspective on each customer contact regardless of the communication channel. Touchpoints compose the customer journey, which can be seen as a process. Due to this, methods of the mature Business Process Management (BPM) can be transferred. One example is the modeling of processes. Our practice-oriented approach has created a sophisticated workshop concept and a modelling notation, which is currently under evaluation in a real-life setting. Furthermore, the research aims to create a better understanding of the touchpoint concept by creating a touchpoint taxonomy for Omni-Channel Management. This will generate benefit for research and practice to intensify integration efforts.