2 Three questions

Integrated knowledge visualization and the enterprise digital twin system for supporting strategic management decision

Purpose

>

This paper proposes an integrated knowledge visualization and digital twin system for supporting strategic management decisions. The concepts and applications of strategic architecture have been illustrated with a concrete real-world case study and decision rules of using the strategic digital twin management decision system (SDMDS) as a more visualized, adaptive and effective model for decision-making.

Design/methodology/approach

>

This paper integrates the concepts of mental and computer models and examines a real case's business operations by applying system dynamics modelling and digital technologies. The enterprise digital twin system with displaying real-world data and simulations for future scenarios demonstrates an improved process of strategic decision-making in the digital age.

Findings

>

The findings reveal that data analytics and the visualized enterprise digital twin system offer better practices for strategic management decisions in the dynamic and constantly changing business world by providing a constant and frequent adjustment on every decision that affects how the business performs over both operational and strategic timescales.

Originality/value

>

In the digital age and dynamic business environment, the proposed strategic architecture and managerial digital twin system converts the existing conceptual models into an advanced operational model. It can facilitate the development of knowledge visualization and become a more adaptive and effective model for supporting real-time management decision-making by dealing with the complicated dependence of constant flow of data input, output and the feedback loop across business units and boundaries.

More

Full text

TranslateLearn more about Translate

Listen

1. Introduction

In the digital age, innovations and effective operations management are critical driving forces to modern business development with digital transformation. The increase of digitization, connectedness and operational integration among different companies and stakeholders in global value chains has transformed business, productions, processes and services systems. There are increasing demands and studies on digital technologies, big data analytics, business analytic skills and their associations with business, management, organizations' dynamic capabilities and performance. From a knowledge-based perspective, a better understanding and analytic skills in the process of strategic management decisions with scenarios analysis would enhance the feasibility of innovation-driven business developments and performance management (Yan, 2018; Yan et al., 2019). In the current digital scenario, organizations are continuing to search for better utilization of data by navigating into an ocean of information from different sources. Since digital transformation is accelerating innovation and forward-looking development of business activities, processes, competencies and models, it is important to fully leverage the changes and opportunities brought by digital technologies and their impact across society in a strategic and prioritized way (Demirkan et al., 2016).

While digital transformation brings opportunities for business excellence, there is a need for the systematic integrations between digital technology, information and strategic decision-making to address business challenges and market demand. Previous studies have proposed the importance of connecting technological innovation, the business model and market as a sustainable innovation ecosystem for industry and enterprise-level knowledge management and practices (Yan, 2015; Yun et al. 2017, 2020; Yan et al. 2018, 2020; Yang and Yan, 2019). Knowledge visualization and representations of information are necessary dimensions supporting knowledge management and decision-making (Sparrow, 1998; Schiuma et al., 2012; Eppler and Brescinani, 2013; Eppler and Burkhard, 2007). Scientific frameworks for visualizing the relationships between knowledge assets and strategic value objectives were proposed (Carlucci and Schiuma, 2007, 2008, 2008; Schiuma et al., 2012). Information visualization can be regarded as a way to create an appropriate visual representation by transforming raw data and information into accessible forms of representations to extract knowledge to viewers, while the importance of applying graphics, diagrams, charts and maps captured in real-time was highlighted (Miah et al., 2017). The approaches and tools of information visualization help decision makers to understand the complexity of their business, their direct and indirect interaction with variety of parties and the environment, enhancing in the assessment of the different trade-offs and the costs associated with the various decision alternatives that can ultimately facilitate the optimization of their decision process and better outcome.

To add a new perspective and practical methodologies for knowledge visualization and to enhance a company's strategic planning capabilities across businesses, a strategic digital twin management decision system (SDMDS) is proposed in this study, with the major characters of the knowledge-based visualized modelling architecture and information technology (IT)-enabled real-time data analysis for dynamic scenarios. It provides a holistic view of a constant and frequent adjustment on every decision affecting the business performance over strategic timescales. The proposed model application not only facilitates the systematical development of knowledge visualization integration for supporting the real-time management decision-making process across different business units but also improves the decision-making process by displaying real-world information and values for every element with simulation analysis to demonstrate the strategic decision-making.

This paper comprises six sections: Section 1 is introduction. Section 2 presents a literature review relevant to knowledge visualization, strategic architecture, digital twin, strategic management decisions, big data analytics, and model-based learning. Section 3 explains the research methodology for modelling works and the case study. Section 4 presents the model applications of the enterprise digital twin system for improving strategic decision-making. Section 5 provides discussions on critical principles and management decision rules for effective applications of the proposed model. Section 6 presents the conclusions and the implications for theory and practices.

2. Literature review

2.1 Knowledge visualization for strategic management decision

Knowledge visualization is to explore the power of visual formats to represent knowledge with aims at supporting cognitive processes in generating, representing, structuring, retrieving, sharing and applying knowledge (Tergan et al., 2006). The application of visual representation will enhance the transfer and creation of knowledge that helps managers, in particular the forms of communication to different stakeholders and multi-objective decision-making problem (Canonico et al., 2021). Visualization techniques play a key role in enhancing the decision-making process and improving the action planning process in the analysis phases (Tan and Platts, 2003). Previous studies also show that with information technology and data analysis, the simulation-based modelling approach has been introduced in both enterprise and industry levels, including new business model development (Eppler et al., 2011), strategic planning (Eppler and Platts, 2009), marketing (Lurie and Mason, 2007), strategic product innovation and dynamic pricing, the knowledge-based management decision support system (Yan, 2017; Yan, 2015; Yan et al., 2019) and entrepreneurial decision-making process (Yan, 2018; Aas and Alaassar, 2018).

Some scholars highlight that visualization’s contents create many opportunities for bringing greater insights to decision makers (McCosker and Wilken, 2014; Miah et al., 2017). It becomes even more relevant in the context of the dynamic business environment and complex system problem. Representation of knowledge is particularly meaningful in the strategic decision-making context where the type, reasons, ways, meaning and scope of knowledge visualization can affect the operation, organizational behaviours and the company governance (Del Giudice and Della Peruta, 2016; Schiuma et al., 2012).

There are advantages and disadvantages of knowledge visualization, including the ease of communications, effective work performances, and knowledge applications and the cognitive, social and emotional risks of visualization from both designer and user perspectives (Bresciani and Eppler, 2015; Troise, 2021). Thus, the growing relevance of knowledge visualization in supporting strategic management decision also calls for a more in-depth investigation of the approaches, models, processes and tools supporting the creation, representation and integration of knowledge at the core of organizational value creation and the business decision-making process (Platts and Tan, 2004).

2.2 Strategic architecture, the decision support system and digital twin

A strategic architecture can be recognized as a roadmap of the future that identifies which core competencies to build and it can also represent invisible intellectual, philosophical and even normative programmes to support virtually all critical business decisions (Kiernan, 1993; Prahalad and Hamel, 1990). By taking modelling techniques and simulations as essential elements in supporting management decisions and operational research, Sterman (2014a) defined management flight simulators (MFSs) as the simulations of complex operational and strategic issues in businesses and other organizations. An MFS is a visualized learning tool that allows managers to compress time and space, experiment with various strategies, and learn from making the rounds of simulated decisions in a designed-learning environment that allows failure and reflection (Bakken et al., 1992). Sterman et al. (2013) proposed that the MFS improves people's mental models in discovering the behavioural patterns of complex systems. When the experimentation is too slow and costly, simulations become the main – perhaps the only way we can discover for ourselves how complex systems work and where high-leverage points may lie (Sterman, 2014b).

A simulation-based strategic decision support system (SSDSS) could be an MFS for supporting the simulations of complex operational and strategic issues in businesses and other organizations (Yan, 2015, 2018). For improving IT-enabled operations management, Yan et al. (2019) proposed a knowledge-based management decision support system (KMDSS) which supports the analysis of the business performance over time and the impacts of strategic resource management decisions on the dynamic capability of e-business from a systematic viewpoint. The visual representation of these models can be regarded as a learning tool and knowledge creation mechanism that allow managers to conduct an experiment with diverse strategies, and learn from simulated decisions in a designed learning environment.

A digital twin is a digital replica of a living or nonliving physical entity, and it refers to a digital replica of potential and actual physical assets (physical twin), devices, systems, processes, places and people that can be used for different purposes (Boje et al., 2020). The digital twin concept started from monitoring high-value, safety-critical pieces of machinery such as aircraft engines in flight and industrial processes. It is an up-to-date and dynamic model of a physical asset or facility that includes all the structured and unstructured information about projects that can be shared among different parties (Deng et al., 2021). In addition, with the application of accurate sensors, data management and artificial intelligence (AI) to the physical objects to collect and transmit data to its virtual replica during the build process, the virtual twin is providing all relevant data about the state of the physical twin (Alizadehsalehi and Yitmen, 2019). The digital twin has become a data-informed model of a physical system and can further identify and predict maintenance issues (Boje et al., 2020). Organization and business performance can be improved by enabling the system to automatically modify its behaviour and have become particularly ubiquitous in the internet of things (IoT) world (Greif et al., 2020).

As aforementioned concepts of strategic architecture, decision support system, and digital twin, strategic modelling for decision-making is essential in the competitive business environment. Simulation-based strategic scenario analysis thus can improve managers and stakeholders’ mental models and make managers’ experience and knowledge more realistic for innovative business development through its continuous real-time feed of data and greater interaction between virtual contents and the real world.

2.3 Big data analytics and model-based learning for dynamic capabilities in the digital age

Big data analytics and dynamic capabilities have been recognized as connected factors for improving organizations' performance over time (Ardito et al., 2019; Shams and Solima, 2019; Singh and Giudice, 2019; Behl, 2020; Medeiros and Macada, 2021). Strategic management with a data driven decision-making culture and quality data has high potential supporting operational and strategic business value creation as it creates actionable ideas for sustainable firm performance and a competitive advantage (McAfee et al., 2012; Gobble, 2013; Wamba et al., 2015; Soto-Acosta et al., 2018; Coluccia et al., 2019; Rialti et al., 2020).

Rialti et al. (2019) proposed to systematize the literature on big data and dynamic capabilities. In addition to managers' experience and intuitive judgements in decision-making processes, information processing and analytic capability also become an essential role in supporting organizational responsiveness and performance, when the industry environment is highly dynamic (Bullini Orlandi and Pierce, 2020). It is worthwhile to integrate the concepts of mental models and computer models with a model-based learning (MBL) system for practical developments of dynamic capabilities in the digital age. MBL refers to human activities interacting with a formal model for systematic learning and cognitive development. It can be a critical part in the digital transformation process from its intuitive experience in management and the evidence-based decision analysis. Mental models are cognitive representations of conceptual and causal interrelations among elements that people use to understand specific phenomena. Computer models are explicit representations of essential parts of reality. With simulation techniques, a computer model provides an environment where a human learner can experiment with hypotheses. Computer models can be of different forms and synonyms such as interactive simulations, microworlds, participatory simulations, interactive learning environments, flight simulators, computational learning devices or the decision support system. Previous studies show that formal computer models, grounded in data and subjected to a wide range of tests, lead to more reliable inferences about dynamics and improve our mental simulations capability (Sterman, 2002; Sterman, 2014a, b). Therefore, utilizations of data analytics and MBL would help to improve dynamic capabilities with the emerging digital technologies.

The recent revolution of information and communication technology (ICT) has pushed the global economic growth as well as e-business performance by an incredibly increasing but fluctuating demand. To the global market dynamics, e-business and Online-to-offline (O2O) commerce with open innovation business models can be a great source of dynamic capability (Yun et al., 2017; Yan et al., 2019). Previous studies also suggested that substantial dynamic capabilities could allow firms to augment and renew resources, reconfiguring them as required to revolutionize and respond to (or generate) changes in the marketplace and in the business environment more extensively (Pisano and Teece, 2007; Teece et al., 1997; Teece, 2017). An integrated approach of knowledge visualization, dynamic capabilities and digital twin systems could facilitate managers with dynamically holistic views and guide towards enhanced prospects for sustainable high performance in a longer time (Teece, 2018).

Figure 1 shows a conceptual model of the business operations and digital twin learning system for decision-making in an organization with an open e-business platform and IT-enabled dynamic operations management system. An open e-business platform normally requires diverse products and innovative visual presentations with different types of IT-enabled services or IT-embedded tasks. Each type of task could be simultaneously supported by the various professional teams, within or across organizations. Therefore, the performance of each team as well as the dynamic operations management will be affected by the working process of related tasks maintained by other teams while their own performance also influences the success of those related tasks.

To assure the strategic alignment between business and IT development, data analysis and information feedback can be continuous flows with the digital twin MBL system and the management. External factors as well as the capability providers and value chain suppliers are also included in the conceptual model as they are critical considerations for strategic planning (Helms and Wright, 1992; Lardo et al., 2020). Previous studies proposed that the system dynamics (SD) methodology has its capacity to deal with accumulations and feedback processes that are well connected to the inherently and dynamically complex characteristic of operations management tasks (Abdel-Hamid, 2011; Yan, 2018; Yan et al., 2019; Yan and Lee, 2021). Simulations and scenario analysis with the SD methodology are powerful analytic techniques to integrate visualization of knowledge creation and the digital twin system for supporting strategic management decisions.

3. Methodology

3.1 Research design and case study

To investigate the integration between the proposed digitalized system and human user-interfaces in theoretical development and practical applications, this research is designed for both quantitative analysis and a qualitative case study. With a simulation model, the case study in this paper consists of a detailed investigation by collecting the empirical materials over a period of time, addressing business challenges and providing an analysis of the context and processes involved in the phenomenon. The phenomenon is not isolated from its context rather is of interest precisely because of the structured-modelling work and the connected business logics and data analysis with the real case. The case study is a team work from identifying the model scope to the modelling development of strategic architecture and simulation-based scenario analysis, working closely with the company for empirical research.

3.1.1 Case selection

This study examines how a UK-based firm makes strategic decisions for business development, how the rationale of decision on human resource management, and how the performance management was enhanced. The case company was established in 2014 with the aim of turning a home improvement in a friendly simple process for the customers. This UK's only assured end-to-end home improvement company provides services from getting a new kitchen or bathroom to renovating an entire building. As home improvement projects are a daunting prospect for many homeowners, the company has created a revolutionary new way to improve customer's experiences. With a team of home improvement experts, designers, surveyors and a network of vetted trusted trade partners, the company designs, manages and builds projects from the start to their accomplishment, and the results are evident in hundreds of successful projects across the UK. Therefore, the successful adaptation to technological and digital advances has fostered its business innovation by capturing the dynamic context of business processes and decision-making that makes this company an insightful case study.

Since the case study with business modelling is not only a purely qualitative research but also a quantitative research that requires empirical data for investigation, data availability is an important consideration in this study. As one of the best practices of model applications, the company uses modelling and management system to pre-test the viability of the venture, to raise substantial funds from investors, and to manage the business with the team that makes it as a representative case in terms of model development and practices.

The case study was carefully selected for certain critical features: (1) it is one of the best case that demonstrates the dynamic process from business start-up, innovation, performance management and strategic decision-making for business challenges; (2) the co-founder and chief executive officer (CEO) of the case has the willingness to work with the researchers that helps empirical investigation and data collection; (3) the case (company) is using the scientific data-driven model as the proposed framework of SDMDS so that real-time data collection and business challenges can be evaluated.

3.1.2 Investigation and data collection

The involvement of the researcher in the process of empirical material generation and interpretation is crucial. For the collection of the empirical material, it is reliable because the researcher knows the case well and directly works with the co-founder and CEO to build specific models and follow up empirical data analysis with the company. This ensures a smooth process and builds a rapport among researchers and participants.

The case study method involves a range of empirical material collection tools in order to answer the research questions with maximum breadth. Based on the model, semi-structured interviews and structured data analysis were conducted along with meeting observations and documents collection. Collecting empirical material from multiple sources allows triangulation (Yin, 2009). This combination of multiple sources of empirical material in a case study method is best understood as a strategy to add rigour, breadth, complexity, richness and depth to the study (Flick et al., 2004). In this paper, the opinions and subjective judgements of the CEO were treated as strategic choices. The data analysis as well as computer simulations are scientific manners to offer objective evaluations for the case study. Both (subjectively) strategic choices and (objectively) data analysis were evaluated with the simulation-based scenario analysis. Therefore, the case study and supporting simulation models have a rigorous process for research findings (multiple methods of data collection to assure the quality of qualitative and quantitative analyses are shown in Figure 2 and Table 1.).

3.2 Research framework

Based on the concepts of SSDSS and KMDSS in the previous studies (Yan, 2018; Yan et al., 2019), the SD modelling technique and digital twin system are adopted for visualizing the strategic architecture and model building. How the managerial digital twin system and strategic architecture facilitate knowledge visualization and decision-making is well explained and illustrated with the demonstrations of a real-world case company.

Digital transformation encompasses the changes taking place in society and industries through the use of digital technologies and drives better performance that has brought to the new concepts and emergent advance technologies (Hess et al., 2016; Majchrzak et al., 2016). One of these new concepts is the “Digital Twin,” which is related to creating a virtual copy of the physical system, providing a connection between the real and virtual systems to collect, analyse and simulate data in the virtual model to improve the performance of the real system and is commonly known as a key enabler for the digital transformation (Kritzinger et al., 2018).

With SD modelling and digital technologies, the characteristics of the stock and flow diagram, casual loop, input and output system make the dynamic business process visualized and realized digital twin system with displaying real-world data to demonstrate the process of strategic decision-making. In our case study, an enterprise digital twin system can be reviewed as a digital replica of potential or actual organizational assets, processes, people, places and systems. It will provide both the elements and the dynamics of how that organization, that department, or that issue operates continually. As shown in Figure 3, there is a correspondence between the organization's real-world elements and the virtual model. The elements presented in a model are matched against things that are observable and measurable in the real world. Instead of picking up real-time data from “sensors”, we are picking it up from the reporting system that every organization operates.

Warren (2005) proposed strategy dynamics approach for analysing a firm's strategy and resource utilization that can further drive business performance. Yan (2018) further proposed the SSDSS framework for resource-based management decision-making that indicates that a firm's strategic outcome as well as business performance could be the results of innovations, business development and resource-based management decisions over time. This study supports a systematic evaluation for time-phase performance with anticipated resources. Figure 4 shows how the digital twin system and strategic architecture for supporting management decisions is enabled by covering the feedback structure of interrelated critical variables, including internal and external resources, actors, dynamic business development, time-phase performance and the continuous and instant management decision over time. With a core strategic architecture, the system can be adaptively modified and extended to cover critical considerations for management decisions.

This paper integrates a standard and rigorous modelling method in a new way with the digital twin concept. There are two features that make it a digital twin model: (1) always displaying real-world values for everything with simulation analysis. This paper suggests that a good model should do what the real world does and for the same reasons we are taking this to the ultimate extreme, (2) the managerial process of using the model to constantly and frequently adjust every decision that affects how the business performs over both operational and strategic timescales. (In the case, the team in the home improvement business got together every week with the model updated with latest data – it would also be easy to do that daily and have the model alert them automatically to any indicator, anywhere in the business, that is going off-track, either bad or good).

4. Case study: applications of the digital twin system for strategic decision-making

First of all, for managing strategic planning and decision-making with the digital twin system, the important distinction is that those physical digital twins pick up every single event that happens to that piece of equipment. It would be an enormous model since the equivalent would be that we would pick up every transaction that happens in the organization, including every person hired and fired, every customer-purchased event, every item ordered and flowed through the supply chain and so on.

The case company's model starts from a very minimal staff-hiring model with short time-units (days). The staffing level requirement is driven by the demand for customer support in the business. This staff group is actually a part of customer-support that leads to a potential digital twin model for staffing management in this demonstration example. The digital twin model could be suitable for operation short-term decision-making and for long-term strategic planning and policy development for the organization. The managerial digital twin requires us to look at processes in some details, and we have to understand what the link is between process map and a dynamic model.

An example of a simple hiring process includes placing an advertisement, screening the applications, issuing job offers, receiving acceptances and checking whether the organizations have got the people they need. If you have the suitable candidate, then you stop hiring and if you have not, you go back and place another job advertisement. In this process, you see that the activities that are being done by the team involved the elements of the system. First of all, they're qualitative and there are no numbers. Secondly, they are “verbs”, they are things that people or groups actually “do”.

We illustrate a matching SD model of “exactly” the same process. As shown in Figure 5, the difference is that you are looking at the stocks and flows of applications received, applications to be reviewed, offers sent out, candidates consideration of the offer, appointments being made. Furthermore, we are accumulating the total number of staffs hired on the right hand side. In the process map, we check if we have got the staff we want, and if we have not, we go back around and place another job advertisement. We can see that the things in the diagram are actually “numbers”. They are quantified, and they are “nouns”; they are actually things that you can see and touch.

In the demonstration of the digital twin model, we are going to change it a little bit. We are going to pick up the same process, but we are going to adapt it for more extensive business challenges. We look at an insurance business with hundreds of thousands of customers. It needs customer support staffs to respond to calls coming in from those customers. A percentage of those customers call for each month and support staff handles those daily calls for each month. If you have understaffing, what you get is overload and the calls are badly handled.

There's always going to be some customer loss rate in a business, but the call overload and badly handled calls would cause the increase of the customer loss rate. Now if you have a staff shortage, what happens in the model is that it triggers a hiring process as shown on Figure 5. Since there is variability in that call rate coming in from customers, we can optionally set a percentage uplift on that staffing target to handle those peak call rates. Furthermore, since there will be episodes that happen in our business, and we can look in particular marketing campaigns to bring in more customers, we may need to impose extra hiring episodes to plan for things that are coming up. Thus, this is the system structure of that staffing challenge. At the top you have got that stock of customers, being won at a rate that is being driven by the marketing campaigns that the organization is running. However, on the right hand side, a number of customers are lost each day. Down at the bottom you have got support staffs who are appointed each day, and there will be some rate of staff turnover depleting that stock of support staffs. The balance between calls coming from customers and capacity of the support staff determines the workload on that team, and if that workload gets too high, then customer support quality goes down and that increases the rate at which customer is lost.

However, the model takes the workload and triggers when the necessary hiring campaigns kick off, and it brings in the applications to exactly the same hiring process as shown in the demonstrative model. Thus, we are going to look at this challenge on two different time scales. The wider time scale is the “strategic” time scale. This business has been running for many years, and it will continue to run for many years. When they are looking at the organization strategy, they will probably be looking back over the last two years. Thus, we were starting at the beginning of 2019, and we are certainly looking out to the end of the two years, that is, 2021. It could easily be a three-year plan going out to 2022 and 2023. Yet the immediate “operational” time window is the current period of time and that is around the period of September to December 2020 (Figure 6). Thus, the person responsible for customer support needs to be looking at what is going on on a daily basis inside that much narrower time window.

As shown in Figure 7, we are looking at an episode which is going to happen shortly, and that episode is the marketing department is going to run a campaign that they believe will bring in 40,000 new customers. We can notice that except for that campaign, the number of customers is actually falling by about 200 per day. It is a bit “more” than 200 per day because an extra number of customers are lost each day when the calls exceed the peak capacity. The reason we would allow it is because it is very expensive to have enough people employed to handle any possible rate of customer calls coming in. Thus, we get the occasional daily spike when there are some bad service episodes and an extra number of customers get lost.

What happens when we trigger that marketing campaign? We will notice that those peaks are more bunched together in that period, so the marketing campaign brings in more customers that we cannot handle. Thus, the customer support quality goes down, and we have a short period of higher customer loss rates. Eventually, the workload drives hiring, and it catches up with that higher call rate, and the customer loss rate falls back to a lower rate. Yet we do not want that higher loss period to happen. As we know this marketing campaign is coming up, we add a short-term hiring campaign in advance of the marketing campaign over on the left-hand side; there is a period where applications are received at a more continuous rate over a number of days. It means that we have got an “in-tray”, if you like, of applications for review in the first stock in the chain.

Then we send out offers to a number of those people, so on any day, there are a number of offers outstanding. It is much higher in the simulated scenario with this extra hiring campaign than it would normally be. That leads to the current number of staff being increased rather quickly over the period before that marketing campaign hits. What that does is it cuts that excess customer loss rate quite sharply, and as a result, we retain more of the customers that have won by the marketing campaign. In practice, every day we add the latest number of calls received from customers. This is a continuing daily operation, every day we look at how many calls have come in and would also probably look at the number of badly handled calls that result from that incoming rate. The next item in the model could track that actual number of badly handled calls, not just the total number of calls.

When we change this from a “short-term” review, what is happening between the total number of calls between September and December 2020 and look at the same issue over that three-year strategic time horizon. The hiring campaign at that particular point in time just looks like a short-term “blip” in that much longer time scale. We can see it compressed, if you like, into a very short episode in the three-year period. It is still there, and furthermore those customers who we retained because of that extra hiring campaign are actually retained into the future and going into 2021 because our regular normal hiring policy ensures that we take on enough staff to mostly look after those extra customers that we have taken on. Notice that we have only done “one” extra hiring campaign in this model. In reality, we would probably do it again for the next marketing campaign, and again for the next one after that, and so on.

Figure 8 shows a whole-business digital twin system of the case company, and the CEO built this model of how he hoped the business should work and he added early data as soon as the business started operating. He shared the model with investors to raise GBP 3 million of new venture funding early in the life of the business. He and his team use the system as visualization of knowledge creation and the digital twin for weekly operating decisions. What they do is provide support for people who want to do home improvement projects. They get homeowner enquiries with web-based marketing and word of mouth. They recruit and validate contractors to do those projects – improve your kitchen or build a roof extension or whatever it may be –and they employ skilled staff to help the homeowners with designing the project, producing drawings, planning the work, looking after the building work itself and handling warranty claims.

Up there at the top is the section of the model that handles the customer growth and losses. In the middle, you have got the call rate coming, in the workload, and the service quality that is triggering that extra loss of customers. Down at the bottom you have got the staff hiring pipeline and the policy that drives that over on the left. The model supports to indicate (1) where you set the marketing campaigns (when it will happen, and there is a little look-up table that indicates how many customers are brought in by each of those marketing campaigns), (2) where you actually enter the updated daily calls-received data, and (3) where you set the staffing uplift that you want to choose to avoid that call overload and where you set the additional hiring efforts. Thus, what we have shown is how we might get from strategy or policy models to digital twin systems for strategic management decisions.

The system can be running on very short time units, even look at maybe hourly events. In this case, you may need to make processes explicit. In a policy-level model, hiring would just be a flow rate in the model. It might be a “delayed” flow rate but that is all it would be. We are looking at the step along that process, the placing of job advertisements, the screening of applications, the issue of offers and the appointment of people who accept those offers. Figure 9 showed a control panel for the business that the management actually uses. At the top you have got the project pipeline – that is how many enquiries per week are coming in, how many projects are currently in planning, how many projects are being completed each week. At the bottom you have got quality and conversion rate indicators, to do with how well enquiries are being handled and how good the actual projects are being done by those contractors. On the right are the financial results, the revenue, costs and profits of the business and on the left are the key decisions to do with hiring staff, seeking contractors and spending marketing to bring in those customers. Each object in this diagram is as the equivalent of a spreadsheet column with its name in the top cell and each period's values in the cells below.

The core pipeline for the home projects business is running weekly over two years, with marketing bringing in those enquires on the left, projects going into planning, projects being managed and that drives the need for staffing. Completed projects drive the revenue, and the impact of quality on the loss of customer enquiries and loss of projects can be evaluated in the planning stage. The quality consequence of an imbalance between projects and the number of staff and contractors can be calculated accordingly.

The above demonstration shows how information visualization by digital twin models for supporting management decisions is able to transform standardized traditional operations and business forecasts into a visualization process that helps an enterprise to establish its unique niche by responding a constant and frequent adjustment on every decision that affects how the business performs over both operational and strategic timescales. The utilization of information and knowledge transferring within organization can further be optimized.

5. Discussions and the principles for improving strategic management decision

The study provides new insights with the integrated knowledge visualization and enterprise digital twin system for supporting management decisions. Strategic architecture facilitates the continuous real-time data flowing into the model and a physical or virtual constant connection that make digital twin system as an integrated multi-physics, multi-scale simulation of a dynamic management decision system. It systematically enables virtual laboratory for business experiments of the individual and team decision-making process by incorporating feedback, delays and non-linearities between variables that support the decision-making environments of executives than merely relying on psychological and subjective judgements in the decision-making process. Consistent with prior studies, knowledge visualization can improve the transfer of knowledge across different parties (Eppler and Burkhard, 2007). It is also understood as both a collective and interactional process and a systematic approach where different players can translate their expertise, share a framework and develop common ground to support decision-making (Canonico, 2021).

In this paper, the major visualization and data analytics contribution is the continuous parallel display of real-world versus simulated results (not only the historic actuals but also the desired-future actuals). In principle, the modellers or managers could always do real-time data analysis and performance evaluation, even with Excel, and the standard management reports often report “latest-month versus budget” for example. Therefore, this paper has demonstrated featured benefits that the real-time data analysis and visualization can be done (1) for every index in the business system and (2) continually updated. It is helpful to support real-world case modelling and applications for improving strategic management decisions. In addition, the proposed models do offer automatic diagramming, in the sense that every change shows up immediately. This paper accepts that businesses pick up data from sensors for minute-to-minute operational control. The innovation is to lift the same principle up to the strategic level and populate the model with business-wide data. Therefore, this paper suggests that it is a “new level of digital twin” applications, with its added value and contributions in advancing knowledge visualization and digital twin management decision analysis.

According to the literature review provided in this paper, previous studies have substantially acknowledged the importance of knowledge visualization, strategic decision analysis, data analytics and computer MBL with digital technologies for improving dynamic capabilities and organizations' performance. To realize these concepts, an integrated knowledge visualization and enterprise digital twin system offers an effective approach to support the strategic management decisions and practical developments of dynamic capabilities in the digital age. Troise (2021) highlighted the main benefits of knowledge visualization in the digital age, including stakeholder engagement, flexibility, knowledge transfer, signalling role, agility and interactivity, while risks such as complexity, absorptive capacity, divergences, capabilities and ineffectiveness are also identified. Therefore, it is important to consider a firm's ability to recognize the value of new information, assimilate it, and apply it to create value and real performance. To improve the effectiveness of model applications, the management decision rules, resource accumulation and implementation strategies can be systematically investigated with numerical data and the information available from the field that help managers adjust and make proper decisions in a timely manner and also improve the knowledge management and organizational learning. This paper summarizes practical implications as well as ten connected principles for facilitating knowledge creation and improving strategic management decision during business innovation as follows:

Make good use of entrepreneurial thinking and innovative practice for business growth: Entrepreneurial thinking replaces the traditional reactive problem-solving management model with “value creation.” Innovative practice processes ranging from enterprise diagnosis, business opportunity exploration, operations management, digital transformation, product and service design, business model development, to market expansion, etc., all of which focus on value creation and performance growth.

Focus on the process for transforming opportunities into real results: We must study for promoting specific programmes and actions for economic or social value creation when faced with all foreseeing opportunities.

Recognize the nature of business dynamics: Regardless of the changes in the market condition, consumer behaviour, technological evolutions, the supply chain and business model, competitor strategies, etc., one must recognize that the core essence of business dynamics lies in value development derived from supply and demand, though such development does not stick to one single form. The life cycle or stage of any product/industry has its share of business opportunities and second curves with open innovations for value creation and market development.

Targeting future developments and performance over time: A good knowledge visualization for strategic development must come hand in hand with a clear-cut definition of the business performance indicator that you pay close attention to, and you must focus on the strategic objectives and performance that might shift over time. It is always necessary to envisage possible development of the future timeline in a transformation or innovation process.

Explain the reasons of changes that occurred in past performance (why)/future goal performance (where)/practical action programme (how): First of all, examine the historical evolution of the past performance that you paid close attention to (why) as your reference experience. Then, develop future scenarios with optimistic and pessimistic estimation where the performance trends might go (where) in the future. And finally, decide what strategic plan should be implemented to support your expected performance in the future (how). These three linked issues can help companies integrate their experiences and strategic objectives, focus on future development, and hatch supportive plans.

Consolidate the critical resources as driving factors of key performance: Attending to the consolidation of critical resources can drive future goal performances. The control on the critical resource helps us reasonably anticipate future development.

Use the stock-and-flow concept for developing corporate assets: Most strategic goals and factor resources are corporate assets that can be continuously accumulated or consumed. Managing flows has significant impacts on the performance of different “stocks”.

Build a strategic architecture to serve as the blueprint for dynamic management: Strategic architecture is a corporate's development blueprint with time-series performance goal for integrated management of various corporate assets' stock-and-flow controls and feedback related systems. In particular, when strategic goals and business/operation systems tend to be diversified, a strategic architecture can greatly visualize and enhance the manager's mental model and vision of resource integration, reduce mistakes caused by partial thinking, and improve decision-making quality and efficiency.

Use computer simulation and data analysis technology for supporting scenarios planning and business development: Diversified scenario planning that assists future business development through data science can promote decision-making analysis efficiency, minimize decision errors from subjective judgement and mental bias, and facilitate cross-discipline communications and group consensus.

Getting feedback from the market and operating data as well as learning from experience and continue to incorporate it into the integration of dynamic management: Lock on the strategic goal of creating corporate values. Integrate experience, data, contingency, expectations and business development by viewing the end as the beginning, coming full circle and start all over again. Systematic innovation and operational optimization are sustainable management capabilities for facing a dynamic environment.

6. Conclusions

The importance of knowledge visualization and digital twin concepts have been emphasized as a source of competitiveness (El Kadiri et al., 2016). Despite some available works, existing architectures and models have rather conceptual characters. This paper demonstrates a scientific approach with integrated knowledge visualization and the enterprise digital twin system by a concrete real-world case study. The decision analysis of the case company is based on a constant and frequent adjustment of every inflow and outflow input and element that affects how the business performs and how the management decision are evolved. This paper explains the visualization process of data, information and behaviours in dynamic business development processes. Both implications derived from this paper for theory and practices are beneficial to further developments.

As the implications for theory, this study improves and advances our knowledge on the emerging research fields in knowledge visualization and the digital twin (Deng et al., 2021). A better understanding of analytic skills in the process of strategic planning with real-time information for current and future scenarios will enhance the visualization of knowledge and realize the feasibility of innovation-driven business developments and performance management. Strategic decision-making is not a single action but are continuous flows through the process of entrepreneurial innovations and operations management. The proposed model helps managers to keep track on the changing of performance through time as well as deal with the complicated additional dependence of its strategic resources flow rates, the constant flow of data input, output and feedback loop for better knowledge management and strategic decisions-making. In addition, a strategic architecture that integrates knowledge visualization with the digital twin system can be seen as an organization's core logic on supporting the decision-making process for sustainable business value creation, and how organizations define and “construct” their own strategic architecture forms an important part of their strategic landscape of business innovation and performance.

As the implications for practices, the case study and business modelling clearly demonstrated that dynamic management can be improved by real-time data input for every element with simulation analysis to demonstrate the strategic decision-making process across different business units and a variety of stakeholders. Since the development of knowledge visualization and the digital twin is still at its infancy as the literature mainly consists of concept papers without concrete case-studies (Kritzinger et al., 2018), this paper illustrates a more adaptive, effective and visualized model for decision-making in the dynamic business world that contributes to relevant business research. To effectively apply the proposed model for supporting management decisions in real practice, ten principles and decision rules are proposed for business applications. In practices, it offers a cost-effective solution with much more confident and comprehensive business decision-making and with less costly manual and administrative works in control. With the vision of digital transformation and business excellence, this paper offers an opportunity to use the model to inform every decision across the business that comprises all existing management reports, such as key performance indicator (KPI) systems, balanced scorecards and other performance management tools in a timely manner.

Note that a knowledge-based decision-making process relies on the decision maker's mental model which has bounded rationality. The model could be limited by the system boundary as well as the variables considered in the strategic architecture. A digital twin model is effective but not able to explain the scenarios driving by excluded variables, therefore, as business model evolves and extends the development of strategic architecture, historical data as well as numerical inputs and the parameters with other critical variables will be required for simulations and scenario analysis consistent with the real-case environment as authentic as possible. For future research, we propose to examine the applications across different industries and incorporate more variables and aspects from human, managerial and environment dimensions for an adequate form of interaction with systematic views to foster entrepreneurial knowledge and dynamic capability.

Ministry of Economic Affairs and Ministry of Science and Technology in Taiwan.

Figure 1

Business operations and the learning system for decision-making in the digital age

[Figure omitted. See PDF]

Figure 2

Modelling with the real case and demonstrative snapshots from interview and analysis videos

[Figure omitted. See PDF]

Figure 3

Visualization of the mental model as strategic architecture for simulation-based decision analysis

[Figure omitted. See PDF]

Figure 4

The digital twin system and strategic architecture for supporting management decisions

[Figure omitted. See PDF]

Figure 5

The structure of the staffing challenge

[Figure omitted. See PDF]

Figure 6

The operational vs. strategic view of customer support staffing

[Figure omitted. See PDF]

Figure 7

The change from a short-term to a strategic view of the same staffing challenge

[Figure omitted. See PDF]

Figure 8

The full model of the digital twin in the case study

[Figure omitted. See PDF]

Figure 9

Demo control-panel of the home-project business start-up

[Figure omitted. See PDF]

Table 1

Methods of data collection and main outputs

Method of data collection Focus Main output

Participant interviews Discussions were based on the role, contribution, interaction with other actors and the process of feedback during the project Model boundary, major variables in the model, strategic architecture, stock and flow diagrams, modelling equations, model parameters

Meeting observations Various aspects such as experience, interaction, participants learning, and so on were observed and analysed in order to map out the value co-creation process Major variables in the model, strategic architecture, stock and flow diagrams, modelling equations, model parameters

Project reports Project reports were key to provide an overview of the whole project, team members' details and history, and the operations of the project Strategic architecture, stock and flow diagrams, modelling equations, model parameters

Meeting notes Meeting notes were used to make sure key points are addressed during the meeting, and it also helped to support field notes taken during meeting observation Strategic architecture, stock and flow diagrams, modelling equations, model parameters

End user feedback documents The process of achieving the feedback and transferring it to other actors is important Strategic architecture, stock and flow diagrams, modelling equations, model parameters, scenario analysis

Data from business operations Numerical data collection from the real business operations Modelling equations, model parameters, simulations, scenario analysis