Abstract
Facing fierce market competitions and fast technological transformation, manufacturing enterprises eagerly aspire to cut down investments in non-core businesses with low returns, such as logistics, warehousing, and machine maintenance. In the context of service-oriented manufacturing (SOM), these non-core businesses are expected to be undertaken by third-party service providers in a centralized product + service manner. Supported by SOM principles, the public warehousing product-service system (PW-PSS) is defined as the integration of public warehouses and related warehousing services (e.g., raw material purchasing, work-in-progress storing/sorting/monitoring, and finished products packaging/distribution) for different manufacturing enterprises in an industrial park (IP). PW-PSS especially benefits small- and medium-sized enterprises (SMEs) who suffer from heavy investment and management burdens of self-built warehouses. To operate the PW-PSS effectively, third-party service providers must improve the service satisfaction degree of customers to attract continuous cooperation. This paper proposes a service satisfaction evaluation (SSE) method considering customer preferences to ensure PW-PSS solutions are more reliable and reasonable.
What Problem Does This Paper Address?
In modern manufacturing, SMEs often face a "warehousing wall." Building and managing a private warehouse requires massive capital investment, specialized labor, and complex information systems—resources that SMEs would rather spend on their core products. While Service-Oriented Manufacturing (SOM) suggests that these non-core tasks should be outsourced to third-party providers in a centralized "product + service" model, a new problem arises: How do you measure if the service is actually good?
Existing evaluation methods often treat all customers the same. However, in an industrial park, one SME might prioritize low cost above all else, while another might desperately need high-speed distribution to survive. This paper addresses the lack of a reliable, preference-driven evaluation method that helps third-party providers design—and SMEs choose—optimal warehousing solutions.
Methods
The authors propose a multi-stage evaluation process that balances objective data with subjective customer needs, as outlined in the five-step framework below:
- Initial Weighting (Standard Impact Loss Method): The system first determines the "initial weight" of different service indexes (like storage space, cooperation time, or distribution speed). This is an objective mathematical approach that calculates how much "significance" is lost if a specific index is not optimized.
- Preference Adjustment (Kano Model): This is the core innovation. The Kano model classifies customer requirements into four categories: Exciting (delighters), Expectation (more is better), Basic (must-haves), and Irrelevant. The initial weights are then adjusted based on these categories to ensure the evaluation reflects what the SMEs actually care about.
- Evaluation (Distance Average Solution Method): Similar to the well-known TOPSIS method, this algorithm ranks solutions. However, instead of just looking at the "best" and "worst" possible cases, it calculates the positive or negative distance of a solution from the average of all available options. This makes the ranking more robust against outliers.
Validation and Results
The paper validates the model using data from an automobile industrial park containing 79 high-tech SMEs. Preferences and satisfaction scores were collected from 50 representative SMEs to populate the Kano model and the evaluation matrix. Domain experts provided the deterministic values used to score eight different warehousing solution alternatives ($A_1$ through $A_8$).
Key Findings: * Preferences Change the Leaderboard: The most striking insight is that a solution that looks "mathematically" superior in a vacuum might actually be a poor choice once you factor in what the customer actually values. With preferences, the ranking was $A_7 > A_6 > A_2 > A_4 > A_8 > A_5 > A_1 > A_3$. Without the Kano model adjustment, solution $A_2$ dropped from 3rd place to last place. * Real-World Accuracy: To verify these results, the authors had the SMEs implement the solutions for 10-day trial periods. The real-world satisfaction scores were "highly in accordance" with the preference-driven model, proving that the Kano-adjusted weights were more accurate than standard objective weights.
Conclusion
The PW-PSS model creates a "win-win" scenario: it cuts costs for SMEs while allowing service providers to achieve economies of scale. While the current model relies on expert scoring for initial values, future work will focus on dynamic scheduling and algorithmic efficiency to handle larger, more complex industrial datasets.
Summary updated