Die digitale Transformation der Gesellschaft birgt enorme Potenziale für Unternehmen aus allen Sektoren. Diese verfügen aufgrund neuer Datenquellen, wachsender Rechenleistung und verbesserter Konnektivität über rasant steigende Datenmengen. Um im digitalen Wandel zu bestehen und Wettbewerbsvorteile in Bezug auf Effizienz und Effektivität heben zu können müssen Unternehmen die verfügbaren Daten nutzen und datengetriebene Entscheidungsprozesse etablieren. Dennoch verwendet die Mehrheit der Firmen lediglich Tools aus dem Bereich „descriptive analytics“ und nur ein kleiner Teil der Unternehmen macht bereits heute von den Möglichkeiten der „predictive analytics“ und „prescriptive analytics“ Gebrauch. Ziel dieser Dissertation, die aus vier inhaltlich abgeschlossenen Teilen besteht, ist es, Einsatzmöglichkeiten von „prescriptive analytics“ zu identifizieren. Da prädiktive Modelle eine wesentliche Voraussetzung für „prescriptive analytics“ sind, thematisieren die ersten beiden Teile dieser Arbeit Verfahren aus dem Bereich „predictive analytics.“ Ausgehend von Verfahren des maschinellen Lernens wird zunächst die Entwicklung eines prädiktiven Modells am Beispiel der Kapazitäts- und Personalplanung bei einem IT-Beratungsunternehmen veranschaulicht. Im Anschluss wird eine Toolbox für Data Science Anwendungen entwickelt. Diese stellt Entscheidungsträgern Richtlinien und bewährte Verfahren für die Modellierung, das Feature Engineering und die Modellinterpretation zur Verfügung. Der Einsatz der Toolbox wird am Beispiel von Daten eines großen deutschen Industrieunternehmens veranschaulicht. Verbesserten Prognosen, die von leistungsfähigen Vorhersagemodellen bereitgestellt werden, erlauben es Entscheidungsträgern in einigen Situationen bessere Entscheidungen zu treffen und auf diese Weise einen Mehrwert zu generieren. In vielen komplexen Entscheidungssituationen ist die Ableitungen von besseren Politiken aus zur Verfügung stehenden Prognosen jedoch oft nicht trivial und erfordert die Entwicklung neuer Planungsalgorithmen. Aus diesem Grund fokussieren sich die letzten beiden Teile dieser Arbeit auf Verfahren aus dem Bereich „prescriptive analytics“. Hierzu wird zunächst analysiert, wie die Vorhersagen prädiktiver Modelle in präskriptive Politiken zur Lösung eines „Optimal Searcher Path Problem“ übersetzt werden können. Trotz beeindruckender Fortschritte in der Forschung im Bereich künstlicher Intelligenz sind die Vorhersagen prädiktiver Modelle auch heute noch mit einer gewissen Unsicherheit behaftet. Der letzte Teil dieser Arbeit schlägt einen präskriptiven Ansatz vor, der diese Unsicherheit berücksichtigt. Insbesondere wird ein datengetriebenes Verfahren für die Einsatzplanung im Außendienst entwickelt. Dieser Ansatz integriert Vorhersagen bezüglich der Erfolgswahrscheinlichkeiten und die Modellqualität des entsprechenden Vorhersagemodells in ein „Team Orienteering Problem.“ / The digital transformation of business and society presents enormous potentials for companies across all sectors. Fueled by massive advances in data generation, computing power, and connectivity, modern organizations have access to gigantic amounts of data. Companies seek to establish data-driven decision cultures to leverage competitive advantages in terms of efficiency and effectiveness. While most companies focus on descriptive tools such as reporting, dashboards, and advanced visualization, only a small fraction already leverages advanced analytics (i.e., predictive and prescriptive analytics) to foster data-driven decision-making today. Therefore, this thesis set out to investigate potential opportunities to leverage prescriptive analytics in four different independent parts. As predictive models are an essential prerequisite for prescriptive analytics, the first two parts of this work focus on predictive analytics. Building on state-of-the-art machine learning techniques, we showcase the development of a predictive model in the context of capacity planning and staffing at an IT consulting company. Subsequently, we focus on predictive analytics applications in the manufacturing sector. More specifically, we present a data science toolbox providing guidelines and best practices for modeling, feature engineering, and model interpretation to manufacturing decision-makers. We showcase the application of this toolbox on a large data-set from a German manufacturing company. Merely using the improved forecasts provided by powerful predictive models enables decision-makers to generate additional business value in some situations. However, many complex tasks require elaborate operational planning procedures. Here, transforming additional information into valuable actions requires new planning algorithms. Therefore, the latter two parts of this thesis focus on prescriptive analytics. To this end, we analyze how prescriptive analytics can be utilized to determine policies for an optimal searcher path problem based on predictive models. While rapid advances in artificial intelligence research boost the predictive power of machine learning models, a model uncertainty remains in most settings. The last part of this work proposes a prescriptive approach that accounts for the fact that predictions are imperfect and that the arising uncertainty needs to be considered. More specifically, it presents a data-driven approach to sales-force scheduling. Based on a large data set, a model to predictive the benefit of additional sales effort is trained. Subsequently, the predictions, as well as the prediction quality, are embedded into the underlying team orienteering problem to determine optimized schedules.

This dissertation investigates the consequences of supply chain disruptions in pharmaceutical supply chains. Across different studies, I examine various impacts of pharmaceutical supply chain disruptions on buyer’s behavior, medication errors, and market share. In Chapter 1, coauthored with Anita Carson, Erin Fox, and Rena Conti, we demonstrate the stockpiling behaviors of buyers during the early phase of the COVID-19 pandemic. Leveraging a quasi-experimental design on IQVIA’s National Sales Perspectives™ data, we show that the sales volume of essential medicines related to U.S. hospital-based COVID-19 treatment concentrated only for the first two months of the pandemic. After these two months, the sales volume of drugs for COVID-19 treatment decreases significantly despite a nationwide increase in COVID-19-related hospitalizations. In Chapter 2, coauthored with Anita Carson and Rena Conti, we examine the impact of a hurricane that decimated the factories of major producers of heparin, an important drug used frequently in hospitals. Using a natural experiment, we find that the hurricane-related pharmaceutical supply chain disruption increased medication error rates of heparin. In addition, we find significant spillover effects. The supply chain disruption increased the medication error rates of a substitute drug. In Chapter 3, coauthored with Anita Carson and Rena Conti, we study how long it takes to recover the market share after the supply chain disruptions using a new metric we propose, Time to Recover Market Share. We explore the differential effects by the brand type of products, the competition level in markets, and the duration of the supply disruptions. With the extensive global supply chain disruptions that we are facing today, understanding their potential consequences is significant. This dissertation advances our understanding of the different impacts of supply chain disruptions and provides practical implications for supply chain members to build resilient supply chains and minimize the effects of supply chain disruptions.

In general companies continuously have to improve their operations to be able to survive in the global competition. They have to be better in utilizing their resources today compared to what they were yesterday. The production systems have changed during the 20th century and factories today do not look like they did hundred years ago. Focus has moved from mass production towards flexibility. The changes in production philosophy have not been followed by a corresponding change in different supporting functions. Research shows that many companies still use accounting methods that have not been developed since the 1930s. The purpose of this licentiate thesis is to provide perspectives on some aspects concerning the relationship between manufacturing operations management and management accounting. This will increase the knowledge and understanding of how management accounting information supports manufacturing decision making. This thesis reports findings from four studies designed to investigate the informational relationship between management accounting and operations management in companies. Results from this research shows that there are many factors to consider when choosing and designing an appropriate management accounting system. Contextual factors include market, manufacturing strategy, technology, and organization. This thesis also reports on the difficulties of making theoretically sound methods work in practice. This thesis contributes with some explanatory aspects on the practical problem and investigates some potential ways forward concerning manufacturing-related management accounting. / Generellt måste företag ständigt förbättra sin verksamhet för att kunna överleva i den globala konkurrensen. De måste bli bättre på att utnyttja sina resurser i dag jämfört med vad de var igår. Produktionssystemen har förändrats under 1900-talet och fabriker i dag ser inte ut som de gjorde för hundra år sedan. Fokus har flyttats från massproduktion till flexibilitet. Förändringarna i produktionsfilosofi har inte följts av en motsvarande förändring i olika stödfunktioner. Forskning visar att många företag fortfarande använder redovisningsmetoder som inte har utvecklats sedan 1930-talet. Syftet med denna licentiatavhandling är att ge perspektiv på vissa aspekter som rör förhållandet mellan produktionsstyrning och ekonomistyrning. Detta kommer att öka kunskapen och förståelsen för hur information ur internredovisning stöder beslutsfattandet inom tillverkning. Denna avhandling rapporterar resultat från fyra studier som syftar till att undersöka informationssambandet mellan produktionsstyrning och ekonomistyrning i företag. Resultat från denna forskning visar att det finns många faktorer att tänka på när man väljer och utformar en lämplig internredovisning. Kontextuella faktorer omfattar bland annat marknaden, produktionsstrategi, teknologi och organisation. Denna avhandling rapporterar också om svårigheterna att få teoretiskt sunda metoder att fungera i praktiken. Denna avhandling bidrar med några förklarande aspekter på det praktiska problemet och undersöker några potentiella vägar framåt för produktionsrelaterad internredovisning.

No description available.

Disasters impact different communities differently due to pre-existing vulnerabilities and inequalities, which diversify amounts and types of public service needs. Understanding the varying needs of communities that rely on government services helps decision-makers allocate limited resources properly during crises to maintain effective, efficient, and equitable service provision across the region. This dissertation includes three independent studies which commonly investigate how service operations can be successfully managed to maintain the operational performance goals of public organizations during slow-onset disasters. The first study focuses on the volatility in the service needs of citizens from a public system during a long-term disaster. The study proposes a time series approach for predicting demand volatility patterns to manage service productivity. This chapter explores the longitudinal impacts of long-term disasters for better service performance management since the timely and accurate prediction of deviations from the expected service demand is vital for utilizing limited resources. The study further discusses the differential impacts of such disasters across locations of socio-economically diverse populations to emphasize the need to consider the diverse needs of people for efficient and effective service provision. The second study builds upon the discussions in the first study and discusses static and dynamic risk factors of slow-onset disasters to reveal how these factors diversify the service needs of communities and impact the corresponding service response performance of public systems during the disaster. The study performs time series analyses to test the impact of capacity adjustments and dynamic disaster risk features on service performance, considering service response time as the performance indicator. The third study focuses on efficient and equitable capacity management and prioritization strategies of an information technology-based public system that experiences significant changes in service demand during disasters. The study presents a mathematical model quantifying the relative service efficiencies associated with service requests from an input-output-based standpoint to uncover the inefficiencies in response performance to different service categories. The paper discusses the opportunities for managing service efficiency and equity within and between service departments by rearranging available capacities and prioritization strategies during emergencies. / Doctor of Philosophy / Disasters impact different communities differently due to pre-existing vulnerabilities and inequalities, which diversify amounts and types of public service needs. Understanding the varying needs of communities that rely on government services helps decision-makers allocate limited resources properly during crises to maintain effective, efficient, and equitable service provision across the region. The three independent studies of the dissertation commonly investigate how service operations can be successfully managed to maintain the operational performance goals of public organizations during slow-onset disasters (e.g., climate change, drought, pandemic). The first study focuses on the variability in the service needs of citizens from a public system during a long-term disaster. The study explores the longitudinal impacts of disasters for better service performance management since the timely and accurate prediction of demand variability is important for resource management. The study further discusses the different impacts of such disasters across locations of socio-economically diverse populations to emphasize the need to consider the diverse needs of people for efficient and effective service provision. The second study discusses disaster risk factors of slow-onset disasters to reveal how these factors affect the service needs of communities and impact the corresponding service response performance of public systems. The study tests the impact of capacity adjustments and disaster risk factors on service performance. The third study focuses on efficient and equitable capacity management and prioritization strategies of a public system that experiences significant changes in service demand during disasters. The study quantifies the relative service efficiencies associated with service requests to uncover the inefficiencies in response performance to different service categories. The paper discusses the opportunities for managing service efficiency and equity within and between service departments.

Yes / This paper focuses on the application and exploitation of Big Data to create competitive advantage. It presents a framework of application areas and how they help the understanding of targeting and scoping specific areas for sustainable improvement. Empirical evidence demonstrates the application of Big Data in practice and tests the framework. An exploratory approach is adopted to the secondary research which examines vendors’ offerings. The empirical research used the case study method. The findings indicate that there is opportunity to create sustainable competitive advantage through the application of big data. However there are social, technological and human consequences that are only now beginning to emerge which need to be addressed if true long-term advantage is to be achieved. The research develops a framework and tests it only in 2 dimensions. This should be expanded. The vendor analysis limitations lie within the nature of the information available and the difficulties in mitigating against bias. The suggested framework can help academics and managers to identify areas of opportunity to do so, setting new levels of performance and new agendas for business. This work contributes to service operations management, building on Kranzberg (1986) and the impact of technology and on Fosso Wamba et al. (2015) by developing a systems application framework to further understanding of big data from a practical perspective to extend their research taxonomy insights. Our case studies demonstrate how the use of BD enhances operational performance.

Yes / This paper brings a new perspective to knowledge by focusing on the application and exploitation of big data in two UK companies providing, respectively, online and branch retailservices. The companies innovatively exploited the data that were generated by new internet technologies to improve business performance. The findings from both case study examples show that benefits do not come simply by adopting technology, but when people think creatively to exploit the potential benefits of ITC. The conclusion drawn is that the realisation of the ‘universal benefits’ of technological innovation does occur, but not necessarily until the hype has subsided. The paper demonstrates that there is opportunity to create sustainable competitive advantage through the application of ITC although the social, technological, and human challenges of managing technology have to be appreciated and managed. These implications need to be appreciated and if true long-term advantage isto be achieved.

Innovation requires an ambidextrous knowledge creation strategy, which is defined as the simultaneous pursuit of both exploration and exploitation. A temporal ambidexterity strategy is one in which an organizational unit dynamically balances its investments in exploration and exploitation over time. This thesis provides new insights on various factors which should be considered when developing and executing a temporal ambidexterity strategy. In the Essay 1, I empirically examine the impact of exploration, exploitation and learning from cumulative innovation experience on the likelihood of successfully versus unsuccessfully generating a breakthrough innovation. The data sample, based on patents in the biomedical device industry, is drawn from the National Bureau of Economic Research patents database. I demonstrate three important tenets for developing a theory of temporal ambidexterity. First, I confirm, as conceptually expected, that when pursued independently, exploration and exploitation have opposing variance-generating versus variance-reducing impacts on innovation performance, respectively. Second, I find that when pursued jointly exploration and exploitation have a negative interaction effect on innovation performance. Third, I show that the benefits of ambidexterity accrue in the long-term, as a result of learning from prior failure experience. However, I demonstrate that prior failure experience and exploitation are jointly necessary, but not independently sufficient, for learning from failure to occur. In Essay 2, I introduce a dynamic optimization model of temporal ambidexterity. I examine the optimal sequencing of exploration and exploitation knowledge creation activities throughout the innovation process. I consider how an innovation manager’s optimal dynamic investments in exploration and exploitation are driven by the innovation team’s knowledge creation capabilities and prior innovation experience, and by the manager’s short-term and long-term innovation risk objectives. The results demonstrate the conditions under which various temporal ambidexterity strategies endogenously arise. Finally, in Essay 3, I extend the single firm model introduced in Essay 2, to develop a model of temporal ambidexterity for two firms jointly pursuing knowledge creation and knowledge-sharing under co-opetition. Here, I consider how co-opetition, that is, cooperative knowledge-sharing with a competitor, impacts a firm’s optimal ambidextrous knowledge creation strategy. I consider two-way knowledge sharing, and I assume that each firm freely reveals its knowledge to its competitor, without receiving compensation. The dynamic analytical results contribute to the open questions regarding optimal knowledge-sharing strategies under co-opetition, by demonstrating under what conditions knowledge-sharing with a co-opetitive partner is beneficial. Importantly, I also analytically examine the factors which drive empirically observed alliance dysfunctions, wherein organizations delay knowledge-sharing and withhold information from their alliance partners.

The concept of ignorance has been unfairly stigmatised in research and practice, and consequently has not received the attention it deserves as a powerful motivator of behaviour in organisations. To understand the role of ignorance, it must be examined as a productive force rather than a shameful weakness, an achievement instead of a failure. This thesis develops an understanding of how ignorance is achieved and why it is perpetuated in the context of managing the Olympic and Commonwealth Games, a series of worldwide mega-events that are popular with proponents of urban development, but which have experienced persistent organisational problems in the form of cost overruns, schedule delays, and scope creep. To do so, this research draws on literature about ignorance from the disciplines of philosophy, anthropology, sociology, and organisational theory, to motivate an embedded case study of Games Organising Committees (OCs) in six host cities around the world. These OCs, which were actively planning the Games during the research, are studied through qualitative research, to develop a dynamic understanding of the role of ignorance in planning the Games. The findings and analysis are presented from two perspectives: the structure of the ‘Games system’ and of the OC; and, the substance of Games planning in the areas of cost, time and scope. While other studies have focused on ignorance as necessary, strategic, and inadvertent, the original contribution to knowledge of this thesis is the proposal of a theoretical framework that focuses on the functional and detrimental outcomes of ignorance. This framework is also shown to be useful in understanding why ignorance persists between organisations, and suggests three basic principles for further research: ignorance as a productive force in management; structure as a scaffold for ignorance; and budget, time and scope as catalysts for ignorance.

A presente tese apresenta proposições para o desenvolvimento e aplicação de técnicas de mineração de textos, de modo a contribuir para a gestão de operações nas áreas médicas e de negócios. Os objetivos desta tese são: (i) identificar e estruturar técnicas de mineração de texto, de modo a elaborar um método para prever internações de pacientes provenientes de emergências hospitalares, tendo como base somente os registros textuais não estruturados escritos por médicos durante o primeiro encontro médico-paciente; (ii) comparar previsões realizadas pelo método proposto no objetivo (i) com análises médicas realizadas por humanos, de modo a verificar se computadores podem atuar de forma autônoma na tarefa de previsão de internações de pacientes provenientes de emergências hospitalares; e (iii) identificar e estruturar técnicas de mineração de texto, de modo a elaborar um método para prever a satisfação de clientes de companhias aéreas, tendo como base as avaliações escritas e publicadas por passageiros na internet. Os métodos propostos utilizaram diferentes técnicas de mineração de textos, sendo validados por estudos de caso. Em relação à área médica, o método proposto pode realizar previsões em tempo real sobre a necessidade de leitos, ajudando as equipes de gerenciamento de leitos a melhorar os processos de fluxo de pacientes. Além disso, verificou-se que tanto médicos (iniciantes ou experientes), quanto máquina, tiveram desempenhos semelhantes na tarefa de previsão de internação de pacientes. Já em relação à área de negócios, o método proposto permitiu extrair dimensões de satisfação de avaliações online, além dos sentimentos associados a elas, considerando diferentes perfis de passageiros, serviços e períodos de tempo. Desta forma, foi possível prever a recomendação de companhias aéreas baseado nas avaliações escritas por passageiros. / This dissertation presents propositions for the development and application of text mining techniques, in order to contribute to operations management in the medical and business areas. The objectives of this dissertation are: (i) identify and structure text mining techniques, in order to propose a method to predict admissions of patients from hospital emergencies, based only on unstructured textual records written by physicians during the first encounter with patients; (ii) compare predictions made by the method proposed in objective (i) with medical analyses carried out by humans, in order to verify if computers can work autonomously in predicting hospitalizations of patients coming from hospital emergencies; and (iii) identify and structure text mining techniques to develop a method for predicting airline customer satisfaction based on online customer reviews. The proposed methods used different text mining techniques, being validated by case studies. Regarding the medical area, the proposed method was able to perform real-time forecasts about the need for beds, helping bed management teams to improve patient flow processes. In addition, it was found that both physicians (novice or experienced) and machine had similar performances in predicting patient hospitalization. In relation to the business area, the proposed method allowed to extract satisfaction dimensions of online customer reviews, as well as sentiments associated to them, considering different profiles of passengers, services and time periods. It also enabled the prediction of airline recommendation based on online customer reviews.