**3. Methods**

To achieve the research goal, the design science research (DSR) method was used. DSR enabled evaluating artifacts from both a user-related and technical perspective [38]. Similarly to action research, the DSR addresses practical challenges while contributing to both practice and theory [39], thus gaining increasing recognition among information systems researchers in the process [40]. It was also successfully used to make designs that provide superior utility in the context of business process management [41]. As the current research delivered a formalization of a pilot EEM implementation process that involved external business partners in a multi-facility organization, the area of application of the method might be considered unconventional. It was, however, fully in line with the systematic literature review delivered by Offermann et al., who identified such artifact types as novel system designs, methods, languages/notations, algorithms, guidelines, requirements, patterns, or metrics [42].

Hevner et al. set forth seven guidelines that enhance the scientific rigor of the DSR approach [43]:


In order to (1) reliably assess the scope of data collected during the pilot EEM implementation process along with (2) determine the technological capability regarding automatic data collection; (3) narrow down the list of activities to those that show the greatest potential in terms of bolstering the EE, and; (4) correctly plan its duration, a party with prior experience in offering EE services was required. Therefore, while developing the artifact, the researchers worked hand-in-hand with the staff of SDC Ltd., a specialized company running engineering activities and related technical consultancy with know-how on deploying telemetry in office buildings and retail facilities. Previous ventures that SDC had engaged in had differed greatly in scale and were not preceded by a standardized preparatory phase that enabled the assessment of the potential energy savings policies. Notwithstanding, their post-implementation reports, project schedules, and work order notes constituted primary data sources while developing the methodology.

The measurement solution used by the company is proprietary. A single control device was installed in the electrical switchboard of each customer's facilities. The device periodically transmitted a JSON frame with the electricity consumption snapshot, along with a list of additional parameters. The development of the artifact required testing the technological capabilities regarding tightening the

interval of transmitting a set of data from the sensors experimentally, as well as analyzing the validity of taking additional parameters into account. It was concluded that the scalability of the analytics required admitting CO2 concentration, humidity, and internal temperature. Datasets were captured by an application implemented in the proven PHP/MySQL tandem. Some of its data processing functionalities were written in Scala. The application processed data and enabled the visualization of electricity consumption. It was possible to generate simple reports. More robust data analyses were performed by a BI solution based on the Qlik Sense engine. The solution took advantage of a data warehouse and enabled multi-dimensional data analyses by combining data from other sources. Those included but were not limited to, physical parameters of the facility, external weather data, and media-related invoices.
