The common problems within maintenance operation management are that maintenance decisions are experience-driven and narrowly-focused. This study proposes a data-driven approach management framework which enables the analysis of dynamic operation data, allowing the management team to modify and improve maintenance operations as well as management decisions. With the development of modern information technologies, a useful data collection system was created, namely, the Mobility System which handles a great deal of useful maintenance operation data for non-residential gas fired appliances in Hong Kong. The implementation of the data-driven management framework was intended to prove the workability of the data-driven approach for maintenance operation improvement and decision making. The generated findings were also reported to the top management as the first phase of enhancement.

Therefore, with respect to the background, it can be seen that the current study would fill two technical gaps. On the one hand, the application of the collected data for the management decision, especially for the management process of the maintenance service. On the other hand, the research work would propose the analytic scheme for the utilization of the collected data so that the organization can enjoy the continuous benefits from the data as well as the analytic framework. Therefore, in order to accomplishing the intention, the research study proposed an evaluative framework for the analysis of the available data with the statistical analysis and tested the proposed framework of the existing data in order to prove the concept for the application of the data. Thus, from the perspective of the organization, apart from the visualization, the proposed analytic approaches provided a strategic opinion for the managers in the decision-making process, while it can be considered an innovation for the industry application.

The overview of the proposed data-driven approach management framework includes three components. The first component is that of data acquisition, which collects operation data from maintenance service orders using the Mobility System, which is then accumulated for data retrieval. The second component comprises data clustering for the management framework and the conducting of a descriptive and segmented analysis. The third component involves the results of the data analysis and the findings for maintenance operation modification and improvement.

The data collection system of the Mobility System aims at collecting essential maintenance operation data, such as material costs, maintenance time used, human resource costs and maintenance quality records, as well as integrating other existing Enterprise Resource Planning (ERP) systems. Following the set-up of the Mobility System for maintenance operations, a vast amount of data was collected and analysed to improve existing maintenance services. After data clustering and data retrieval conducted by the Mobility System, the data analysis focuses on the maintenance operation comprising six stages, namely, the plan, execution, reporting, analysis, improvement and standardization. The performance of each stage was measured by the deviation between the management target and the actual performance of four aspects, including human resource allocation, spent cost, time used and delivered quality. In terms of the demographic information of the maintenance operation, this includes the type of maintenance, type of gas-fired appliance and the scale of the project. By reading the ranks of the four key performance indicators, as well as their comparisons, the deviation between the actual conduct and the expectation of the management team can be depicted. The segmented analysis, based on the variance analysis, revealed the diversity of the performance and expectation with regard to the various kinds of operations and their related processes. All these were essential to the rationale and decision making of the management team.

In addition to the management framework, this study also utilized the data to generate the findings for the company. Three data-driven findings were proposed to modify and improve the maintenance services. Having presented and discussed the results of the findings with the management of the company, it was agreed that the data-driven approach management framework is an innovative way of modifying and improving the maintenance operation. The implementation of the pilot study findings was launched subsequently.

Firstly, due to the needs of organizational learning and the updates of the specification, we aim to improve the process of standardization. The deviation between the management target and actual performance was found in spent cost, human resource allocation and delivered quality. A follow-up investigation was conducted to investigate the situation. Secondly, small projects demonstrated excellent performances while the review on the allocation of resources was proposed. Finally, this study shows that the processes have shared performances from a statistical viewpoint and suggests the simplification of the evaluative scheme.

Overall, this shows that the data-driven approach for the modification and improvement of the maintenance operation was accomplished successfully in terms of the proposed maintenance management framework and the practical implementation of the scheme.