Business Cases for Digital Twins in Biopharmaceutical Manufacturing—Market Overview, Stakeholders, Technologies in 2025 and Beyond

Abstract

Over the past years, the biopharmaceutical industry has been marked by substantial growth, with key players dominating market sales. A major change in research and development is the onset of digital twin (DT) technology in manufacturing. This work gives an overview of the market and major stakeholders, such as contract development and manufacturing organizations (CDMOs), regulatory bodies, and academia, their functions, and challenges. Fundamental concepts and definitions are reviewed and serve as an overview of the challenges ahead of the full adaptation of DTs in manufacturing. Using established market analysis tools, the environment is analyzed, and a business case is developed. Opportunities and threats for small startups and larger pharmaceutical companies to gain a competitive edge are analyzed and evaluated. Even small groups of 4–16 employees enable significant margins at a return on investment of less than 1 year.

1. The State of the Industry

A digital twin (DT) refers specifically to a comprehensive digital representation of a physical object (the physical twin) that is capable of bidirectional communication with that object [1]. DTs enable improved operational efficiency by optimizing the manufacturing processes in real time. This can lead to reduced downtime, waste, and resource consumption, and batch-failure reduction [2]. This efficiency translates into cost savings [3] through lower cost-of-goods and faster time-to-market for their products [4], aligning with the operational excellence discipline and thereby maintaining their competitive advantage.

This paper presents existing stakeholders of the industry in regard to DT for biopharma manufacturing implementation. Existing guidance and guidelines, concepts, and misconceptions are discussed, to then demonstrate tools and different strategic models are demonstrated, which enable stakeholders to find business cases.

The biopharmaceutical industry generated in the year 2023 a total prescription drug sales volume of 1053 billion USD [5], while the OTC (over-the-counter) market was valued at around 200 billion USD. The top 10 companies accounted for more than 44% of these sales, while the top 50 had a share of more than 80% [6]. It is clear that a few key players in the industry have significant influence over the direction the sector moves and, among others, what technologies, like the DTs, will enter the market.

Of course, besides these originator or innovator companies, there are more stakeholders involved. Especially with regard to the manufacturing business, contract development and/or manufacturing organizations (CDMO), as well as control system and automation solution vendors, are stakeholders as well. Their business is at least partially, for control system and automation providers, but often completely, in the case of CDMOs, based on the business of the originator companies.

In 2023, the top 10 contract development and manufacturing organizations (CDMOs), ranked by sales, generated a combined revenue of 30 billion USD. Despite this being considerably smaller than the revenue of originator companies, CDMOs play a crucial role in driving innovation, particularly in areas such as digitalization, automation, and the implementation of digital twins. Their need to remain competitive and attract clients drives them to lead advancements in cutting-edge technologies like digital twins.

The other major stakeholder group in digital twin development comprises regulatory authorities such as the U.S. Food and Drug Administration (FDA) and the European Medicines Agency (EMA). These organizations are vital to the biopharmaceutical industry because they provide guidance on good manufacturing practice (GMP) and quality standards to ensure safety and efficacy. They also offer critical guidelines for computerized systems validation (CSV) [7,8], which are integral in implementing advanced technologies. However, to our knowledge, there is currently no specific guideline dedicated to, specifically digital twins. Despite this, the principles provided by these authorities, such as those concerning data integrity, are crucial in guiding the exploration and adoption of digital twin technologies within the pharma sector.

In addition to regulatory authorities such as the FDA and EMA, the International Society for Pharmaceutical Engineering (ISPE) plays a significant role in the context of digital twin development. Although not a regulatory body, ISPE is a major contributor to the creation and updating of guidelines that provide comprehensive directions for good automated manufacturing practice (GAMP) [9]. Their efforts are important in addressing the evolving questions raised by the industry [10]. Moreover, ISPE facilitates valuable opportunities for discussion and collaboration by hosting conferences and events focused on these topics, providing a platform where industry, regulatory, and academia can share insights and advancements. Another important platform is the BioPhorum. This industry-exclusive consortium has different working groups on pharmaceutical areas of interest. In the context of DT implementation, the development working group, especially the workstreams in silico modeling and PAT, monitoring, and control, publishes frequently whitepapers that give additional guidance [11,12,13,14,15].

The last major stakeholder group in the field of digital twin development is academia. As the cornerstone of modern digital twin manufacturing and technology, academic institutions are responsible for providing essential process models [16,17,18] through engineering and mathematical principles from disciplines such as chemical engineering, computer science, applied mathematics, and statistics. Developments like the transition from batch to continuous processes [2,19], innovations in chromatography [20,21,22,23,24], water-based processes [25,26,27], and process intensification [28] have been heavily driven by academic research, as evidenced by the extensive literature on these topics in areas like machine-learning and soft sensors [29,30], debottlenecking and scale-up [31], drug solubility [32] and solid drug formulations [33,34]. In contrast, the industry’s contribution and adaptation to these advancements occur at a slower pace, as reflected by the relatively low number of existing continuous manufacturing processes compared to traditional batch methods. Additionally, academia is a crucial source of subject matter experts (SMEs), whose expertise is both limited and invaluable.

Above all stakeholders, there is the common socio-political goal to make medicines more affordable, while also keeping the environmental impact of industrial manufacturing minimal. The affordability can be obviously correlated to the cost-of-goods (COG) of manufacturing, and the environmental impact is typically expressed in global warming potential (GWP in kg_CO2/kg_Product). There are also other indicators used in the biomanufacturing community, like the process-mass-intensity (PMI in kg_Waste/kg_Product), but they are unsuited to reflect how the usage of a green solvent or water based process has less impact on the environment and sustainability goals compared to toxic, organic and emission intensive solvents [35,36,37].

Figure 1 shows the GWP over COG for industrial products. Bulk chemical industry with base chemicals to exemplify such as urea and ammonia (MP methanol pyrolysis or WE water electrolysis) or methanol (EHC electrolysis hydrogen and carbon dioxide or MP methanol pyrolysis) can be found at the very corner of minimal GWP and COG, older and newer green processes as well [38]. This is the result of more than a hundred years of process improvements and efficiency developments culminating in by now continuous, automatic, and to some degree autonomous production and early digital twin adoption. This was driven by fierce competition, high market demands in terms of robust supply, but also affordability and quality.

Of course, the scale of production capacities is in the range of millions MT/year, so there is also economy-of-scale benefiting these sectors, however this alone should not satisfactorily explain the magnitudes of higher COG and GWP the biomanufacturing industry currently operates at (note in Figure 1 the double log-scale for GWP and COG). This is a known issue, regulatory industries have therefore for decades now encouraged the industry to move to continuous production, based on quality by design (QbD)-based process design and control strategies enabled by the adoption of modern process analytical technology (PAT) equipment and moving towards the digital technologies, e.g., digital twins [39]. According to ICH Q8, QbD is a systematic approach to the manufacturing process development that begins with predefined objectives and emphasizes product and process understanding and process control, based on sound science and quality risk management [40]. QbD addresses regulatory concerns by enabling control strategies that link critical process parameters (CPP) to critical quality attributes (CQA). These QbD-based process development and control strategies and their link to DTs for biomanufacturing have been demonstrated by several academic and industrial working groups [9,41]. The application of QbD principles on DT development and validation is discussed by Uhl and Udugama [42,43] and illustrated in Figure 2.

Biologics dose is low in µg to mg amounts due to high efficacy molecules; therefore, the typical production capacities needed are the lowest compared to other branches. A transition of bulk chemicals in the lower left edge to the higher right area, biologics are the natural extract products for pharmaceuticals, as botanicals. As a large-scale benchmark for the sugar refineries based on sugar cane are shown which are nearer to the world-scale bulk chemical manufacturing efficiency. The academia community has contributed to all these fields in research and development, publishing and discussing these technologies, concepts, and their potential benefits in GWP and COG reduction in conferences and workshops, for decades and internationally [44,45].

The slow adoption of modern manufacturing standards by other, more efficient industries should not be just linked to higher demands in terms of regulation and quality standards and larger molecular product size. It is a common phrase in the biopharma space that “for new products everyone wants to be the first, but for new technologies everybody wants to be second”. This mindset is not adequate to provide responsible answers to our common socio-political goals of fighting global climate change, while also enabling fair and affordable access to the best medicines. It does, however, highlight a certain skepticism towards these, at least in the biopharma space, new technologies, which can often be traced back to a lack of basic knowledge, which leads to less imagination ability for innovations. So, the following will provide some insight into digital twin technology, concepts, and verification and validation methodologies, to demystify the digital twin as a use business case for a gain in all three management goals of profit and product quality, and risk minimization. Furthermore, company guidelines and CEO/management gratification are more and more related to GWP reduction towards the goal of climate neutrality as well.

Industry has made notable progress in networking and contributing to publicly available knowledge in the realm of digital twins. One example of this is the modeling workshop series, which held its fifth iteration in 2023 in Denmark [46]. This series provides a comprehensive overview of the overall status and chronological progression of various modeling applications from an industry view. By engaging with these initiatives, the industry demonstrates a commitment to advancing collective understanding and fostering collaboration, thereby enhancing the synergy between academic research and practical industry application.

However, there are several aspects that were mentioned in [46] that we would like to address:

• Over the past decade, the industry has come to recognize a positive return on investment (ROI) in modeling and, by extension, digital twin technology, which is a positive development;
• There is, however, still a need for clear definitions as they are crucial for clarity and precision and are still frequently misused:
  • As stated right in the introduction, a digital twin refers specifically to a comprehensive digital representation of a physical object (the physical twin) that is capable of bidirectional communication with that object [1]. This definition is common and is also prevalent in the process automation community, where it includes modifying the behavior of a manufacturing asset in response to its physical state. Udugama et al. suggest a five-step implementation strategy, from simple balance equations to fully validated process models, that allows real-time process optimization through model-based advanced process control [42,47];
  • A digital shadow (DS), meanwhile, involves a validated, mechanistic model with at least one real-time interface for process data acquisition [48,49,50]. However, it lacks a direct feedback interface to the physical process, limiting its use for real-time advanced process control with direct feedback loops [51,52];
  • Mechanistic models, also known as physico-chemical or rigorous models, form the foundation of DT and DS and are precisely defined to distinguish the effects of different phenomena such as fluid dynamics, phase equilibrium, and mass transfer. Properly considering fluid dynamics differences enables effective scale-up from laboratory to manufacturing scale;
  • Distinguishing between mechanistic, hybrid, and data-driven models is useful. However, care must be taken to prevent overlap in statistical hybrid models, which could compromise the basic mechanistic structure, rendering them purely data-driven;
  • Model calibration refers to the process of determining model parameters using experimental data. In the biopharmaceutical industry, there appears to be a lack of familiarity with fundamental chemical engineering practices, which have established these concepts, including quantitative model validation, for decades;
  • Plant modeling, often referred to as flowsheet simulations, routinely involves cost estimations and scheduling, as seen in tools like SuperPro and Aspen Plus. Flowsheet modeling tools also include software such as gPROMS, which is more process model-focused.
• Implementing a digital shadow as a recommender system with no direct control on the process can be an effective initial step in introducing technology and increasing familiarity [3,53]. The usefulness of a digital shadow is demonstrated in [54], where real-time data during a chromatography run enhanced model prediction so that precise differentiation between two overlapping components was possible;
• The estimated requirement for 3–4 subject matter experts must consider the number of projects, workload, and interdisciplinary expertise needed:
  • Process engineering for modeling, experimental plans for model validation, and parameter determination;
  • Mathematics and informatics for sensor interfaces, models, process control systems, and statistics for PAT applications and data management;
  • Analytical knowledge for validating inline PAT with offline QA (quality assurance) methods;
  • Understanding the QbD approach, including risk analysis and robust statistical methodologies;
  • Regulatory considerations require each type of expert to be represented by three individuals to account for absences and ensure continuity.
• Regulatory approval follows established guidelines, with process models categorized as type III high-level models that influence product quality. Model validation with small-scale experiments, followed by validation runs comparing outcomes with and without digital twins, is a feasible approach.

Further comments on the 5th modeling workshop include several observations. Model parameters can often be efficiently determined within a few weeks using established methods. They are the basis of quantitative validation of models that emphasize prediction uncertainty through experimental parameters rather than focusing solely on the models themselves. It is also important to note that data-driven models are generally not suitable for quality-related decisions, and current literature on these topics is frequently neglected. While there seems to be a perception of ambiguity in regulatory workflows, clear guidelines are indeed available to guide the process effectively.

Nevertheless, common sense allows us to point out that digital shadows are a support system for operators. The operator finally decides within the binding approved operation parameter ranges. Therefore, no additional validation is necessary. Digital Twins are of model type 3 with product quality relevant decisions; therefore, statistically sound three additional validation batches to validate process performance with and without digital twins for autonomous process operation seem to be logical for such manageable efforts.

2. The Automation Pyramid

The automation pyramid (Figure 3) is a conceptual model that organizes the various layers of automation and control systems within a manufacturing facility, from the most basic to the most integrated [55,56,57].

The 0 level consists of field devices such as sensors and actuators. Digital twins at this level involve real-time data acquisition and process monitoring of individual unit operations or equipment, ensuring accurate digital representation of the physical components. The operational time scale here ranges from milliseconds to seconds, depending on the unit operation and control task.

Level 1 incorporates programmable logic controllers (PLCs) and distributed control systems (DCS). This level manages process control at the equipment and unit operation level. Digital twins provide real-time supervisory control, monitoring, and optimization of specific processes, allowing for efficient fine-tuning. The time scale typically spans from milliseconds to minutes, again dependent on unit operation and control task [58].

Level 2 manages supervisory control and data acquisition (SCADA) systems, it oversees data gathering from the control levels. Digital twins offer oversight across multiple unit operations, facilitating process optimization and data aggregation for extensive analysis. The operational time scale also extends from seconds to minutes [43,59,60,61].

Level 3 is connected to the so-called manufacturing execution systems (MES), and handles production workflows, scheduling, and quality control. Digital twins provide holistic simulations for broader material, energy, and inventory balances across processes. Typically, this is the boundary from pure manufacturing processes digital twins to broader plant digital twins, that do not necessarily need to rely on mechanistic models. They would rather combine the information from levels 0–2 digital twins and calculate this information in terms of broader flow balances. They support decision-making and operational strategy, operating on a time scale of minutes to days [62,63].

Finally, level 4 is often managed by enterprise resource planning (ERP) systems. ERP system software integrates business and financial management processes, like finance and accounting, and supply chain management. Digital twins simulate logistics, supply chain dynamics, and inventory management, offering strategic insights and aligning manufacturing with business goals. ERP systems are not to be confused with the mechanistic digital twins from levels 0–2. The operational time scale here ranges from hours to days, sometimes extending to months.

The purpose of illustrating the interconnection of automation pyramid levels with digital twins is to highlight that there are indeed existing guidelines that provide insight into validation and qualification for computerized systems utilized for manufacturing, and therefore, should be considered when validating and qualifying a digital twin.

3. Digital Twin Validation

A computer system within a company operating in a regulated business encompasses more than just hardware and software. It also includes any equipment and instruments connected to the system, as well as the users interacting with the system or equipment according to standard operating procedures (SOPs) and manuals [64].

Computerized system validation (CSV) is a documented procedure mandated by regulatory agencies globally to ensure that a computerized system performs its intended functions consistently and reliably [65].

It is therefore part of the common technical document triangle (Figure 4), in specific module 3—quality, which summarizes the application parts of new molecular entities (NME) and biologics license applications (BLA).

We have already established that the digital twin of a manufacturing process covers levels 0–2 of the automation pyramid. It is evident that it needs to be evaluated just as any SCADA/DCS system would be, as it is crucial to product quality and offers similar functionality in terms of remote management, automation, alarm indication, data transmission, and user access and security. The basic guidelines covering these validation requirements are laid out in GMP guidelines, specifically Appendices 11 and 15, as well as ICH M4Q and ICH Q8-10, and must be followed strictly. There are various standards that complement these frameworks, including but not limited to the ISO family (ISO 9000, ISO 27000, ISO/IEC 12207:2017), ISA standards (ISA 18, 88, 101, and ISA 112), and most prominently, the ISPE GAMP5 guidance (A Risk-based Approach to Compliant GxP Computerized Systems). The FDA recently published a new guidance on considerations for the use of artificial intelligence to support regulatory decision-making for drug and biological products. They propose a risk-based credibility assessment framework. The multistep approach (steps 1 to 7) can be followed to establish and evaluate the credibility of a model. Although this guidance covers artificial intelligence models, the adoption of this for mechanistic, hybrid, and data-driven models is reasonable. In this context, FDA scientists also published recently an examination of process models and model risk frameworks for pharmaceutical manufacturing, applying also the risk-based framework [67]. The presented cases cover lyophilization as well as chromatography and PAT models. This is particularly useful, as they demonstrate how to perform steps 1 to 7 and define model risk in terms of model influence and decision consequence, which are criteria adopted from the V&V40 standards [53]. The level of regulatory burden associated with a submission and lifecycle management will depend on its potential impact on product control and quality [14]. As the model’s influence on product quality increases, so does the regulatory expectation for validation and documentation. The levels are categorized into low, medium, and high impact models [68]. To support long-term goals and facilitate effective use, companies may decide to increase validation and documentation requirements earlier in the development process. The digital twin with communication channels of data and information between the manufacturing process and digital instance [1] would certainly fall under category 5 in the ISPE system, meaning it has the highest demands in terms of knowledge and technical expertise (Figure 5). These are, however, part of the business model of typical control and automation solution providers such as Siemens and Körber Pharma, and not a hindrance to digital twin adoption. Körber Pharma does offer support in validation and verification of DT along V&V40 standards [53]. More on this validation and verification standard for computational models can be found at ASME (American Society of Mechanical Engineers) [69].

Many excellent overviews have been published covering the details of these norms and guidelines, such as those by the QbD-Group. As a quality affecting computer systems, these norms are no different for a digital twin than they are for a SCADA/DCS system. However, what is unique to a digital twin, and may lead to hesitancy in adoption, are the mechanistic models at its foundation. Particularly, the sensitivity, accuracy, and precision of these mechanistic models are often still met with concern. This mainly reflects the slow adoption and acknowledgment by the industry of foundational research from academia.

Distinct and quantitative verification and validation workflows have been published for several years now for different unit operations and modalities (Figure 6). Over the past years, these have been expanded to include further validation as digital twins, covering key qualification aspects such as the definition of critical process parameters for design and operating spaces, validation of PAT strategies and interfaces for real-time feedback loops, and control strategy development and testing.

There are now several industry studies demonstrating the ability to effectively adopt digital twins as part of the manufacturing control strategy, proven by engineering runs [60].

4. Business Case and Models

The previous sections are meant as an introduction to not only stakeholders, but also to define the current state in digital twin technology and existing frameworks. The following sections are now dedicated to exemplifying how the business case for digital twins can be assessed from different stakeholders’ perspectives.

4.1. The Value Discipline Model

The value disciplines model (Figure 7), developed by Michael Treacy and Fred Wiersema, categorizes companies based on their focus into three primary value disciplines: operational excellence, product leadership, and customer intimacy. It is an excellent model to explain the different motivations a pharmaceutical company can have by moving towards digital twin technology. Each discipline represents a strategy that can help businesses achieve a competitive advantage. Applying this model, pharmaceutical companies can leverage digital twins as follows:

As already stated in the introduction, DTs enable improved operational efficiency by optimizing the manufacturing processes in real-time. This can lead to reduced downtime, waste, and resource consumption, and batch-failure reduction [2]. This efficiency translates into cost savings [3] through lower cost-of-goods and faster time-to-market for their products [4], aligning with the operational excellence discipline and thereby maintaining their competitive advantage.

In regard to product leadership, in the pharmaceutical industry, product innovation and quality are paramount. Digital twins facilitate sophisticated R&D (research and development) by allowing detailed modeling and simulation of drug formulations and manufacturing processes. This can lead to faster advancements and higher-quality products, reinforcing the product leadership discipline. By adopting digital twins, pharmaceutical companies can stay at the forefront of innovation, bringing cutting-edge therapies to market more rapidly and setting industry standards.

In regard to customer intimacy, digital twins can enable the faster market readiness of innovative drugs and vaccines. Through enhanced process performance, digital twins can help to enable treatments to individual patient needs (personalized medicines), improving efficacy and patient outcomes. Real-time monitoring of drug manufacturing allows for rapid optimization and quality control. This ensures that drugs and vaccines are not only innovatively developed with cutting-edge technology but also meet the highest safety standards.

Efficient processes lead to cost savings in materials, time, and labor, which makes high-quality drugs affordable at lower prices. This affordability is crucial for broadening access to necessary medications, particularly in low-income regions [72]. In particular, during pandemics, the ability to quickly develop and distribute effective vaccines and treatments is paramount. Digital twins enable rapid prototyping and scaling-up of vaccine production [73], which is critical in responding swiftly to emerging health crises. This technological advantage allows for the expedited deployment of vaccines, ensuring that populations have timely access to life-saving vaccines.

In the following, further methods of analysis were used in the development of a business case, which serve to analyze the environment, the industry, the company, and the project. The models are exemplified from the position of a start-up positioning itself and its business strategy towards providing digital twin solutions. The methods are, however, universal. Existing businesses as well as manufacturing companies, which are treated in the following as customers, can adopt these methods from their perspective.

Analysis methods used for the strategy development and implementation of a business model are the following:

Figure 7. Value disciplines model. Adapted from Yokogawa [74].

Figure 7. Value disciplines model. Adapted from Yokogawa [74].

4.2. Porters 5 Forces Model

The 5 Forces Model by Michael E. Porter, presented in ‘Competitive Strategy: Techniques for Analyzing Industries and Competitors’ [75], is used to analyze the competitive forces within an industry and to assess the intensity of competition and profitability. The model identifies five central forces which, in their interdependence, represent significant influencing factors.

The first force is the threat posed by new competitors, which depends on the barriers to entry. High barriers, such as strong regulatory requirements, can reduce the threat of new market entrants [75]. The second force, the bargaining power of suppliers, depends on the number of suppliers and the availability of specialized materials, which increases the risk for buyers. The bargaining power of buyers is the third force and refers to the ability of buyers to control price and quality. Buyers with high purchasing power can exert pressure on suppliers [75]. The threat of substitute products is the fourth force, which increases price pressure and can affect the profitability of companies. Finally, rivalry among existing competitors describes the intensity of competition. High rivalry can lead to price wars and aggressive marketing strategies [75].

Overall, the 5 Forces Model provides a valuable basis for assessing the competitive landscape of companies. It enables managers to identify relevant external forces and adjust their strategic planning to achieve competitive advantage and ensure long-term profitability [75].

4.3. Growth-Share Matrix (BCG Matrix)

The Boston Consulting Group’s Growth Share Matrix (BCG Matrix) is a strategic planning method that helps diversified companies allocate their resources (Figure 8).

In visual terms, the BCG matrix is represented as a two-by-two matrix, where the two dimensions are market share and market growth rate. The matrix can be used to sort and categorize different products into so-called stars, cows, dogs, and question marks. General rules can be found for these categories. Milk the cows, sell the dogs, invest in the stars, and analyze the question marks.

4.4. McKinsey Matrix

The McKinsey matrix is also a model used to evaluate strategic business units (SBUs) and is essentially a revised version of the BCG matrix. It is based on two dimensions, market attractiveness and competitive strength. Market attractiveness replaces market growth in the BCG matrix. It therefore contains more factors that can be used to determine the degree of attractiveness of a product. At the same time, the dimension of competitive strength replaces market share, which determines the strengths of a product in addition to the company’s market share. The McKinsey matrix has three evaluation levels (high, medium, low) and therefore allows for more nuances in the results compared to the two levels in the BCG matrix.

The strategies in the McKinsey matrix (Figure 9) can also be seen in recent developments. Emerson acquired 55% of AspenTech shares in 2022 for 6 bil. USD [76] and recently finalized its merger with the acquisitions of the remaining shares for 7.2 bil. USD [77]. This strategic move aligns with other consolidations in the digital twin industry, like the acquisition of CD-adapco and PSE by Siemens in 2016 [78] and 2019 [79] to strengthen their portfolio with the StarCCM+ and gPROMS software suite. Another example is the announced partnership of Rockwell Automation and DataHow [80].

4.5. 7-S-Model (McKinsey)

The basic premise of the model is that there are seven internal aspects of an organization that must be aligned if it is to be successful. Since the variables are interrelated, no significant progress can be made in one area unless corresponding progress is made in other areas.

The seven interdependent elements of the McKinsey 7S Framework are visualized in Figure 10. All the factors in the McKinsey 7S model are equally important; however, these seven factors are further divided into two categories: strong and weak elements. The first three elements, which are strategy, structure, and systems, are strong elements because they influence the management in a company. On the other hand, the last four, namely shared values, skills, style, and staff, are the weak elements as they are culture-oriented and more abstract. McKinsey places shared values at the center of this model. Shared values basically include norms and behaviors that are expected of all employees.

4.6. Value Chain Analysis Method (Porter)

A value chain is a series of activities that a company performs to produce a product or service. In product development, this includes the journey of a product from conception to final sale and service after the sale, as well as everything in between, such as the procurement of raw materials, manufacturing, marketing, etc.

Porter divides the activities into two categories: ‘Primary Activities’ and ‘Secondary Activities’. The key to a successful value chain analysis is to find out which processes have problems and then implement solutions in due time (Figure 11). In addition, the individual activities are rated from strong to weak influence on a scale of 0 to 4 [81].

5. Results

The first step is to analyze the current situation. This includes analyzing the environment, the industry, and the company itself.

5.1. Analysis Environment

The economic environment has a decisive influence on an industry, a company, and a management program. The social and cultural environment can also be important for both multinational and domestic companies.

Potential customers for the implementation of a DT are pharmaceutical companies. These are listed in the following in Figure 12 according to their Rx (prescription drug) sales and their headquarters’ locations, based on the Pharmaceutical Executive from 2024 [6]. In total, the top 50 result in a total Rx sale of USD 868,242 billion. A large proportion of the particularly high-revenue companies are located in the AMER (America) region. However, 41.66% of Rx sales are generated in the EMEA (Europe, Middle East, and Africa) region. APAC (Asia Pacific) accounts for 12.33%, the lowest ratio of Rx sales.

5.2. Analysis Industry

The second phase involves analysis of the industry in which the firm operates. This phase can be critical, particularly in terms of how the firm’s product is defined. A clear differentiation is made between the potential market and the company’s own competitors, listing the 5 forces and ranking them according to their significance. A low ratio, therefore, represents a low threat.

Porter’s five forces model comprises five elements (suppliers, new competitors, customers, substitute products, industry competition) that a company can use to analyze the attractiveness of a market (Figure 13). As a classic analysis tool, the five forces model can be used in almost any form of market environment analysis relating to an industry, a company, a business segment, or a product. The elements of the five forces model enable a holistic view of the current and potential future market environment.

Force 1—Threat of new entry: The threat from new competitors is considered to be low, as it is unlikely that new companies will offer DTs. To be able to enter the market, it is essential to have sufficient knowledge of the research and development of DTs in bioprocesses. This is only an option for potential university start-ups that work in cooperation with larger software companies. Force 2—Threat of Substitution: The focus here is on substitutes that offer the same function for the customer, although they may differ. This concerns less innovative options, those of established manufacturing concepts that rely on manual operation and offline testing. Experience has shown that pharmaceutical companies often do not care if the COGS can be improved. Force 3—Supplier Power: Suppliers represent a rather minor influencing factor in this case. As this is a digital product in the form of software. The only factor of influence is qualified staff. Due to the general decline in the number of students in STEM fields, there will be fewer potential staff available in the future. Force 4—Buyer Power: The bargaining power of customers is one of the two most powerful factors in this case. The benefit here is the high number of pharmaceutical companies that have rarely used digital twins up to now. This is possible due to regulatory requirements, which mean additional work for process validation after the integration of a DT.

Already established competitors represent a highly competitive risk, especially in the future. The high ratio of large and well-known companies such as shown in Figure 14, which often offer complete packages, makes it difficult for newcomers to enter this industry.

Figure 14. Competitive strength vs. market attractiveness of individual products. The bubble size corresponds to sales (the criteria used for evaluation by authors are given in

Table 1. Typical factors that influence the competitive strength and market attractiveness.

Competitive Strength	Market Attractiveness
Absolute position	Market profitability
Relative position	Market growth
Global market access	Porter 5F: Threat of new entry
Products	Porter 5F: Threat of Substitution
Value chain	Porter 5F: Supplier Power
Innovation	Porter 5F: Buyer Power
Economic position	Porter 5F: Competitive rivalry

).

Figure 14. Competitive strength vs. market attractiveness of individual products. The bubble size corresponds to sales (the criteria used for evaluation by authors are given in Table 1).

Companies that offer a competitive or complementary product portfolio around digital twin technology in this scenario are shown in Figure 14.

In the competitive analysis, based on the usual subjective evaluation of the authors, potential companies and their product portfolios were arranged in a matrix consisting of an x-axis, which represents market attractiveness, and a y-axis, which represents competitive strength. The size of the bubbles corresponds to the market volume. The color coding also shows the product categories in which the individual products are active. It is possible for a product to cover several areas. The categorization in the matrix is based on a purely qualitative assessment. Products such as SIMATIC from Siemens and AspenOne are in a particularly favorable area at the top right, which indicates their predominant competitive strength and high market attractiveness. These companies are highly likely to be able to gain significant market shares and further expand their positions.

In the top left region, products such as StarCCM++ from Siemens and DataHowLab show moderate competitive strength, but in a less attractive market segment. This indicates that they may struggle to grow on their own in the long term and can be tackled by strategic investments or mergers [78,80]. An overview of established competitors with product portfolio is given in

Table 2. Established competitors with a product portfolio.

Company	Relevant Software Products
Siemens	Star-CCM++, gPROMS
Körber Pharma	PAS-X Savvy
Cytiva	UNICORN™, GoSilico™
DataHow	DataHowLab
Novasign	Hybrid Modeling Toolbox
YPSO-FACTO	YPSO Proxima®, YPSO-Ionic®
Rockwell	MES-Software-PharmaSuite
ABB	ABB suite, ABB Ability™ Expert Optimizer
Securecell	Lucullus®
Emerson	DeltaV™
AspenTech	AspenONE Engineering software

.

Products that show high market attractiveness but have a weaker competitive position may face the challenge of improving their strategies to increase their competitive strength and capitalize on the attractiveness of the market. Companies in the lower left-hand range that are weakly positioned in terms of both market attractiveness and competitive strength may need to consider fundamental strategic changes to become more competitive in a changing market environment.

5.3. Analysis Company

In this part, the focus is on analyzing the company itself, not only in comparison with the industry and in comparison with the industry average, but also internally in the form of quantitative and qualitative data. This phase includes factors such as objectives, constraints, management philosophy, strengths and weaknesses, and the structure of the company.

When analyzing the most important competitors, the strengths and weaknesses are examined in comparison to the competitive environment. This results in a strategic intention and various conclusions. Established companies already have good access to the market and a different customer base, depending on the size of the company. Furthermore, many companies, such as Siemens, offer a broad product portfolio, designed to make it easier for customers to integrate new software systems. In most cases, the complexity of the products remains very low. Hybrid and data-driven systems are not offered. Instead, fixed operating points are part of many offers.

At present, to the knowledge of the authors, there is no process that is authorized with a DT that actively intervenes in the manufacturing process.

5.3.1. Results of Value Chain Analysis (Porter)

Primary activities are critical to creating value and competitive advantages. These include inbound logistics, production and development, marketing and sales, outbound logistics, and service (Figure 15).

Inbound logistics focuses on efficient procurement and the provision of the necessary software tools. Only suitable, non-proprietary software tools are used. The lack of vendor lock-in results in a low level of dependency. In development, the focus is on agile development methods and optimized coding and testing processes. The focus is on the continuous development of DTs for biopharmaceutical applications. This includes laboratory tests, simulations, analyses, etc. In this scenario, development plays an important role in the value chain as it forms the basis of our DTs. The better the development, the more mature the final product.

Marketing and sales focus on strategic advertising campaigns and effective distribution channels, while service emphasizes regular updates and responsive customer support. The sales organization has a high degree of influence as it is in direct contact with customers and has a significant impact on the external image. Outbound logistics focuses on warehousing, shipping, and improving delivery times through automation and continuous integration, which have little impact on the value chain as a digital product does not need to be stored or shipped. Services are the activities that maintain the product and enhance the customer experience. This includes additional services such as training, warranties, refunds, and regular updates. In particular, staff training and models, and software updates are often requested and required by customers. They, therefore, have a significant impact on the value chain.

The task of the secondary activities is to support the primary activities more efficiently. Increasing one of the four secondary activities helps at least one primary activity work more efficiently. These include the areas of corporate infrastructure, personnel management, technological development, and logistics. The corporate infrastructure is defined by a clear management style described in the 7S model. The organization is funded through research projects with industry partners. Human resource management includes the ability to recruit, train, and retain employees. Our business case is about training potential students/employees to become experts needed for the development of high-quality DTs. With the number of students in STEM programs declining, this small number of students needs to be managed in the best possible way. Technology development is about incorporating cutting-edge technologies and continuous innovation. This includes continuous research and development of new control strategies, but also the further development of the increasing understanding of processes. This is reflected in the number of papers published in this area. In logistics, the focus is on the careful selection of development tools and licenses in conjunction with cost-saving strategies. Work equipment, such as laboratories and offices, must be available. This is the only way to ensure smooth development.

After defining the primary and secondary activities in the value chain, strengths and weaknesses can be uncovered, but also clear statements can be made about possible strategies. The value chain analysis shows that the divisions have different added values. This leads to poor value creation in the individual areas with a low number of employees, as the resources cannot be optimally distributed to the individual areas, and a cost disadvantage arises from the activities. This corresponds to the start-up phase of a company with few employees. As the number of employees increases, value creation can increase by allocating more staff to the high value-added areas, such as production and development, marketing and sales, and service. The activities create a competitive advantage.

5.3.2. BCG Market Growth–Share Matrix

The BCG matrix illustrates the various development phases that a product can go through during its life cycle (Figure 16). At the beginning, during the product launch, this state is classified as a ‘question mark’, as it is only possible to determine whether a viable market exists for the product during the course of its market presence. Provided that an adequate marketing strategy is implemented, the product has the opportunity to develop into a ‘star’ and, thus, achieve a significant market position.

It is crucial that the ‘star’ is kept on the market through targeted marketing and regular maintenance, and services until it is saturated. After this phase, the product is transferred to the ‘cash cow’ segment. In this phase, it is particularly important to secure the product’s long-term market presence. Strategies for this include regular software updates, staff training, and the implementation of subscription models that integrate such services.

Products that are categorized as ‘poor dogs’ should be avoided, as they have both a low market share and low growth prospects. Minimizing these types of products is of central importance for the long-term competitiveness of a company.

5.3.3. Return on Investment Studies

After the strategic business case analysis as presented above, an ROI analysis was performed to evaluate potential growth scenarios. 3 different product packages are defined to illustrate and exemplify the analysis:

• “DS package”, which is a digital shadow package to implement at first a DS to get used to the technology, engineering, and consultant efforts, is either directly paid without any license strategy or with a license strategy;
• “DS package+” is the service above, but with an additional quality by design (QbD) control strategy without any PAT implemented, but a dedicated process control space, again with and without any license strategy options;
• “DT package” the final digital twin for model predictive control in order to automate operation, including the PAT approach under QbD, and 3 additional batches for validation of such

Within any market growth analysis, different percentages of demand for the above 3 products are evaluated (Figure 17). The outcomes for each scenario are as follows:

−

Scenario 1: No Market Growth

• In this scenario, the group begins with 4 employees, with staffing doubling after 4 and 7 years. There is no initial funding, and the first product sale occurs in year 4. Product distribution is 20% DS digital shadow Package, 60% DS Package+ (plus), and 20% DT digital twin Package, focusing on “DS Package+” due to market interest. The sale of five licenses annually results in a profit from the first year, with a Weighted Average Cost of Capital (WACC) of 6.5%.

−

Scenario 2: Market Growth and Decline of Product Type Demands

• This scenario considers a 20% growth in “DS Package+”, a 40% growth in “DT Package”, but a 20% decline in “DS Package”. Profit is achieved from the first year, with results showing a 13.08% to 20.8% increase in annual revenue compared to Scenario 1.

−

Scenario 3: Significant Growth for “DT Package”

• Here, “DT Package” experiences 100% growth, while sales of “DS Package” and “DS Package+” remain steady. Profit increases with each year, assuming regulatory approval for digital twins, making it a realistic scenario attractive to the industry. This results in a 17.87% to 28.63% increase in annual revenue compared to Scenario 2.

This aligns with other studies, e.g., by DataHow, who found assuming a DT caused process efficiency gain of 50%, an ROI of 131 million USD, and if freed process development resources are invested into expanding the development funnel of up to 1.25 bil. USD. These ROI assume up to 10 blockbusters, with 12.5 bil. USD sales in total [82].

The industry claims a shortage of subject-matter experts in process modeling and digital twin implementation [46]. Contrarily, there is a substantial international academic community actively working in this domain, frequently publishing research and producing young professionals with exactly this expertise. However, a scarcity of job opportunities in this area has led many to pursue job positions in other biomanufacturing fields, such as MSAT (manufacturing science analytics and technology), with a focus on traditional process development rather than the more innovative DT approach. Additionally, the limited available positions with a DT focus often require extensive industry experience, compelling recent graduates to specialize elsewhere, exacerbating the issue. This situation appears to be a self-inflicted challenge, particularly within the biopharmaceutical sector.

6. Conclusions

There is a clear business case for digital twins. As can be seen by upcoming start-ups and the presented analysis, the push towards Pharma 4.0 is ongoing, and early adopters of this technology will have a competitive advantage. The business case demonstrates that even a small team of 4–7 core employees can effectively develop and sell digital twin solutions. By extension to about 16 employees, this approach is even more feasible for larger pharmaceutical companies, which can benefit significantly from establishing an in-house digital twin team. In-house groups do have the benefit of reliability, consistency, and references for risk minimization in contrast to start-ups, who have to gain that at first to be taken seriously as contractors. Such initiatives would not only enhance operational efficiencies and innovation but also align with strategic goals for growth and market leadership in the evolving pharmaceutical landscape. Pursuing these technologies now positions companies to thrive in a future increasingly shaped by digital transformation for the added value of further product and supply safety.