Multi-Channel-Beziehungsmanagement für den elektronischen Kaufprozess

Andreas Hilbert, Stefan SommerFKBI 2009

Introduction

As retailers transitioned to multi-channel models in the mid-2000s, a significant analytical gap emerged: the "known stranger" paradox. While a customer might have a long history with a physical store, they often appear as a complete stranger when first visiting the company's online shop. This lack of data integration leads to poor personalization and inefficient recommendation engines. This paper proposes a systematic framework to bridge this gap by integrating customer relationship management (CRM) across stationary and electronic channels.

Theoretical Framework: Customer Lifecycle Mapping

The core of the proposed framework is the adaptation of the classic customer lifecycle into a multi-channel context. Traditionally, the lifecycle is viewed through phases of acquisition, retention, and recovery, each requiring different management strategies.

The classic customer lifecycle model integrated with management phases.

The authors refine this model into five specific phases tailored for the electronic purchase process:

  1. Emergence (Entstehung): Potential customers or anonymous visitors.
  2. Development (Entwicklung): New online customers who may possess an extensive offline transaction history.
  3. Penetration (Durchdringung): Established online customers with growing digital footprints.
  4. Diversification: Recovering former offline customers through online channels.
  5. Focusing: High-intensity, stable online relationships.
The adapted lifecycle model for the electronic purchase process, detailing phases from emergence to focusing.

Optimizing Personalization via Cross-Channel Weighting

The paper addresses the "cold start" problem by utilizing historical offline data to inform online recommendations. The authors argue that the "Development" phase is the most critical juncture for data integration. By weighting offline purchase history heavily (e.g., an 80/20 split) during this initial phase, retailers can provide relevant suggestions immediately.

As the relationship matures into the "Focusing" phase, the system should dynamically shift its weighting toward more recent and contextually relevant online behavior. This lifecycle-dependent approach ensures that personalization remains accurate as the customer's digital profile grows.

Algorithmic Considerations

The choice of recommendation algorithm should also depend on the lifecycle phase. For the "Emergence" phase, where only session data is available, User-based Nearest Neighbor filtering is appropriate. Once a user is identified and transaction history (offline or online) is available, Item-based Nearest Neighbor filtering becomes superior for precision.

Conclusion

Integrating offline data into the electronic buying process significantly reduces information deficits in early relationship stages. While the proposed weighting factors are currently theoretical, they provide a robust roadmap for future empirical validation. Expanding this model to include additional channels like mobile commerce remains a key area for future research.

Summary updated

BibTeX
@inproceedings{hilbert2009multi,
  author = {Andreas Hilbert and Stefan Sommer},
  title = {{Multi-Channel-Beziehungsmanagement für den elektronischen Kaufprozess}},
  booktitle = {FKBI},
  year = {2009},
}
Figure