Summary:
The main goal of this study was first to present how daily sales could be collected by markets and then to evaluate if the use of daily sales for a certain product category could improve the forecast accuracy of the following three months. First a step-by-step process has been created in order to allow PMI Duty Free (PMIDF) markets to select the best process to enter daily sales into the PMIDF database. Then this study applied the lean six sigma methodology to the daily data collection process to allow a dynamic reporting of daily data using Tableau. Finally this study evaluated if the use of daily data allows to improve the sales forecast accuracy.
Results:
The time required by daily sales to be displayed in the dynamic dashboard has been reduced by 77 % by applying the lean six sigma methodology. The use of daily sales allows to significantly improve the sales forecast for a given location.
Résumé:
Dans cette thèse sont analysés les concepts importants de la gestion de la production, appliqués à un cas particulier. Le périmètre de l’étude est limité pour seulement inclure la théorie qui est pertinente dans notre cas d’étude. L’entreprise étudiée est une manufacture horlogère de luxe.
Disposer d’un service client de qualité est devenu, pour beaucoup d’entreprises, une priorité. L’entreprise, de par sa position stratégique, se doit d’avoir un service client irréprochable en matière de qualité et de délai.
Ce projet de master s’est efforcé d’analyser en détail le flux complet de réparation dans l’entreprise, dans le but de proposer des améliorations pour diminuer les temps de passage, tout en respectant la qualité exigée par la manufacture.
Résultats:
Afin de disposer d’un flux de réparation efficace, une implantation a été planifiée, intégrant les propositions d’améliorations faites au cours de ce projet. Cette implantation servira de base pour préparer le déménagement qui aura lieu dans les prochains mois.
Une de ces propositions, qui consistait à créer un îlot de diagnostic, a été implantée et a permis de réduire le temps de passage du diagnostic d’environ 30%.
Summary:
As the role of data in business grow each day, more enterprises, especially multinational ones, shift their ways to become more data-driven. This shift causes a great necessity for new management disciplines to overcome data related challenges within enterprises. For a number of imperatives, Master Data Management (MDM) discipline is one solution to this need for enterprises to be able to properly manage and govern its core business entities. The aim of this study is to determine and present several incentives which are deduced through a study performed in PMI, for deployment of MDM within a multinational enterprise.
Results:
By exploring master data challenges and issues within PMI, the incentives for multinational enterprises can be briefly presented on three levels as following. On the individual level, supporting employees to become more efficient and less frustrated with their daily responsibilities and activities involving master data is the main incentive. While on the departmental level, an increased intra- and inter-team productivity with reduced overheads makes MDM deployment attractive for multinational enterprises. Moreover, better and stronger sub-organizational strategies on the departmental level with improved insights gathered through an MDM practice presents another incentive. Finally, all of the above compound and translate into the enterprise level incentives for multinational organizations. Last but not least, creating harmonization across core business functions, having more accurate, reliable external financial reporting capabilities and enablement of data-driven strategies with factual decision-making mechanisms provided with a successful MDM practice are the final high-level incentives for multinational corporations to deploy MDM.
Summary:
This thesis studies on how to improve the accuracy for real-time demand forecasting. There are two topics discussed in the thesis: 1) how to improve univariate time series forecasting accuracy by upgrading Merck’s demand forecasting tool and applying pattern recognition method; 2) how to improve multivariate time series forecasting accuracy by selecting the optimal predictor subset.
Results:
Three methods are developed to effectively improve demand forecasting in real business world.
The first method is to upgrade Merck’s demand forecasting tool – SAP IBP by optimizing the configuration of forecasting algorithms. The principal contributions of this updating process are:
- Test the forecasting performance of new forecasting algorithms added into SAP IBP, such as gradient boosting decision trees (GB). It is found that GB gets lower forecasting MAPE (Mean absolute percentage error) than linear regression methods in non-stable products.
- Determine the optimal parameters setting of forecasting algorithms aiming to obtain the lowest MAPE.
- Identify the optimal combination of algorithms to gain lowest MAPE.
- Improve forecasting performance by lowing MAPE by 12.01%. Increase the weighted SFA (sales forecasting accuracy) by 4.5%.
- The upgraded SAP IBP forecasting model came into use in June of 2019 at Merck.
The second method proposed in the thesis is time series pattern recognition method by applying clustering. The main contributions of this pattern recognition method are:
- Develop an innovative time series representation method by combining SMA (simple moving average) and PAA (piecewise aggregate approximation) to overcome the natural constrains of time series data. This method enables time series data to adapt to the classic algorithms such as clustering, which provides high quality of shape-based pattern detection ability.
- Develop a pattern recognition method which can efficiently filter out the groups of products sharing similar trend changed patterns.
- Decrease forecasting MAPE of trend changed products by 31.22% by cutting down the historical training dataset.
The third method presented in the thesis is to improve time series forecasting performance by applying multivariate time series forecasting. The major achievements of this method are:
- Prove the time series representation method proposed in the thesis is a customized method. Users can fine tune the parameters of this method according to their objectives. By adjusting the parameter, time series data can be proper adapted to the liner regression subset selection methods, such as best-subset selection, forward stepwise selection and shrinkage method.
- Fill in gaps of time series predictors subset selection methods.
- Decrease forecasting MAPE by 10.43% by applying best 3 predictors selected by subset selection methods.
Summary:
The current project took place in the Direct Materials (DIM) Procurement department of Philip Morris International (PMI). The objective of the project was the implementation of a Global Information System that will provide visibility to DIM Procurement leadership on Key Strategic and Operational Performance Indicators, ensure standardization of metrics and build a data-driven organization. The project had a global scope involving information from all the regions where PMI operates. Its development was achieved through the standardization of data and processes across regions, workflow automation and the development of a Dashboard by using Tableau.
Results:
- Identification and standardization of different processes across regions based on a set of best practices which were defined centrally.
- Creation of new roles and responsibilities within the Direct Materials Procurement Organization concerning the proper supervision and governance of the system.
- Automation of the workflow within the scope of the system, reducing the need for manual interventions and improving efficiency.
- Development of a Global Key Performance Indicators Dashboard which significantly reduces the time and complexity of information retrieval and improves decision making with data-driven insights that were previously unavailable
Summary:
The emergence of the Internet of Things (IoT) has generated new opportunities for many businesses. In particular, IoT can be a source of meaningful data about the customers of a company. So, many firms believe that IoT can be used with CRM (Customer Relationship Management), a business that intends to optimize the relationships of companies with their clients. That’s for instance the opinion of the American research firm Gartner.
Therefore, this thesis investigated how IoT can be used to enhance CRM. The thesis is composed of a theoretical study and a practical study. The theoretical study reviews the Business around IoT and the technical side of IoT. To rely on a concrete use case, the practical study was made with a Swiss Energy Supplier: Groupe E. An end-to-end proof-of-concept was developed from this use case. It includes IoT devices and a CRM software.
Relying on this experience, the thesis provides new insights on challenges, issues and possibilities to associate CRM and IoT. Notably, it investigates how data can be retrieved, stored, processed and finally used to enhance the CRM. The PoC makes use of other useful resources, such as a Cloud platform, a business analytics software, a low-code platform to develop business applications, and Machine Learning tools.
Results:
IoT can be used in many ways to enhance the CRM. But some challenges need to be overcome to build a successful project such as: protecting against security threats, ensuring the quality of data, monetizing an IoT business case. In addition to IoT devices and to the CRM software, many other resources should be adopted to develop an end-to-end solution. For the infrastructure of the solution, the main questions concern the choice of devices, the choice of network, the use of some Edge computing. To manage the data, a Cloud platform such as Microsoft Azure can be employed. It allows to retrieve data from the devices, to store it, to process it with powerful algorithms, and to make it usable for the end-user. Machine Learning can bring a lot to the processing of the data. Embedding resources in one another allows to monitor the whole solution in a single dashboard. It improves a lot the user experience. All these tools contribute to enhance the CRM with IoT.
Summary:
This project is a thorough analysis of the registered production data of the production lines in the ETA factory in Sion. The goal was to create a work environment supervised by the experience of these online data collections realized daily by the information system in place. The ERP, MES, and the HR database, which represent all the information system of the company, were tackled to assess their reach and use their potential. The analysis focuses on the different aspects of time embedded in the system and checks the flow of that information. Finally, based on the analysis and the critical milestones of the analysis, different tools were developed to, first automate the analyses and then to measure some key indicators on the line — a close eye was kept to work with a Lean management environment to match the company policy.
The analysis focuses on one pilot production line, which is studied in depth. All the results are about this line but all the tools and indicators built are generalized to the whole production structure.
Results:
The first step of the analysis is a reproduction and a check on all the measurements and calculations of the systems. One significant type error on the measure of personnel time was spotted, and the different analyses concluded that they are due to the MES precision. A solution to bypass the problem is then introduced which corrected the irregularities. All the other problem of the measurement and the line organization are listed in the analysis and correction are suggested.
The report concludes by presenting the tools and indicators developed after the analysis. A workable database has been built for the HR department, and an indicator for the personnel production time is in place to spot and correct irregularities. A general tool has been put in place to allow managers to visualize the data from the ERP and have an accessible overview of the production lines and their productivity. Finally, based on these automated tools and their measures of productivity, a sketch of a tool to measure the need in personnel and assess the factory needs in personnel is presented.
Summary:
For many years, Pampers led innovation in the baby care industry and still holds the position of global leader on the market. However, revenues began to decline as Pampers failed to adjust to millennial parents’ needs in time. In an effort to secure its market share, an independent team was created and organized like a startup to promote lean innovation within the company. The purpose of this project was to develop a 10 years innovation roadmap to enter the connected baby care industry. A market analysis was conducted to investigate current competitors and customers’ needs, and resulted in the identification of 12 customers’ pain points. Benefit spaces were prioritized within the 10 years to structure a consistently growing new revenue stream of $2 billion. Finally, the lean startup methodology was studied to support a successful market entrance.
Results:
12 customers pain points were defined: diaper weting, sleep, teething, feeding & nutrition, stool diagnostics, motor and cognitive developments, personalized skincare, relationship between caregivers and the baby, air quality, noise, light and temperature.
The latter were prioritized within a 10 years plan according to potential revenues, demand and complexity of implementation
A financial model was built showing the evolution of the revenues for the next 10 years
The study of the lean startup methodology revealed that Pampers managed to reduce its usual product time to market, however the team still follows many Pampers processes. Hence the time to market could be even more optimized if the brand decided to spin-off the branch dedicated to connected baby care.
Résumé:
Les composants nécessaires à l’assemblage des bracelets ROLEX sont conditionnés pour permettre leur distribution aux ateliers. Néanmoins, on constate que les capacités choisies pour les conditionnements ne sont pas toujours adéquates, et l’activité de préparation et conditionnement des composants impacte directement la production. L’objectif du stage est de proposer une méthode d’optimisation des tailles de lots et de la mise en godets des composants bracelet. Un outil d’optimisation actualisable et applicable aux autres sites de production devra être fourni. Le rapport détaille l’application d’une démarche de Lean Management aux procédés de mise en godets et de préparation des composants bracelet chez ROLEX SA, dans le but d’optimiser et rationnaliser les tailles des lots conditionnés. Il décrit la création d’un outil capable de déterminer automatiquement les tailles de lots optimales pour chaque composant. Sont exposées également diverses propositions d’amélioration et de gestion de la mise en godets. Les scénarios sont chiffrés en termes de gains afin de déterminer quelle est l’option la plus prometteuse, le tout en considérant les contraintes techniques, ergonomiques et humaines posées par les ateliers d’assemblage et la mise en godets. Pour le scénario le plus optimal, une étude financière plus approfondie est menée afin de comparer le coût d’un éventuel changement du système actuel.
Résultats:
- Gain mensuel de près de 500 heures de préparation et mise en godets.
- Réduction de 17,7% du nombre de godets produits quotidiennement.
3. Production d’un outil d’analyse et d’optimisation actualisable.
Summary:
Product proliferation has emerged as a natural response to the market and its evolving requirements. Consequently, the increase in product diversity is often related with higher profits and executed through different strategies. However, it can also lead to increasing operational costs due to generated complexity within the supply chain, arising from having broader product portfolios. One common purpose of performing complexity cost analysis is to identify actions that can potentially reduce operational costs and increase margins. Several methods exist to estimate the economic impact of product variety management. These methods include approaches ranging from activity-based to product-based considerations, among others. In this thesis, the core elements and drivers related to variety-induced complexity are identified and converged into a framework, named CCF, capable of measuring the complexity costs from an operations viewpoint, considering the impact from different functional departments of a company. The method focuses on upstream and downstream activities of the supply chain which are triggered not only by usual logistics operations, like inventory management and transportation, but also related to commercial or marketing activities linked specifically to product variety management. The latter are normally neglected in traditional methods. An important factor for developing the method is the cost-to-serve approach, which is normally performed separately instead of being a part of an integral cost analysis. The design of the framework is presented in a way that it can serve and help companies interested in analysing their complexity costs within their operations. The framework is evaluated using a company in the Fast-Moving Consumer Goods (FMCG) industry as an example. By presenting a practical example of how the framework is applied, the method and its components are explained within a specific context. Additionally, mitigation strategies to reduce complexity costs, implemented in several companies, are presented and analysed.
Results:
The results validate the method by providing coherent costs that are aligned with the expectations of where the complexity, generated from product variety, is more likely to impact the bottom line. It is hoped that this thesis will inform companies about variety-induced complexity management and to assist them in their complexity costs assessments.
Summary:
“You will either step forward into growth or you will step back into safety.” once said Abraham Maslow.
The goal of this thesis is to develop a blueprint for Satisco’s development that can be replicated in any European country.
After a presentation of the company, I developed a conceptual framework on global expansion based on Alain Verbeke’s work and included it in a general decision-making process. I then created a hands-on methodology for managers to apply during the launching phase in an international context.
Based on these elements, the thesis provides strategic recommendations for Satisco’s future global expansion.
Results:
Based on theoretical content as well as on my internship experience as business manager for Satisco Switzerland, I created two blueprints for Satisco’s international development.
The first blueprint provides guidelines to managers for decision-making about where to expand. The first step is to conduct a full market study to determine the customers’ demand, the candidates’ offer and the competitive environment. Then, the enriched theoretical framework can be applied. I advice managers to conduct in parallel a feasibility study to support the final decision-making.
The second blueprint, directly withdrawn from my internship experience, provides an abroad development path for business managers during the launching phase. In particular, the Market Opportunity Navigator has been presented as an efficient tool to pilot the strategy in a fast-changing environment.
Based on these two blueprints, the thesis provides strategic recommendations for Satisco’s future global expansion.
Summary:
Machine Learning (ML) projects are novel for many businesses and little research has been done on their execution. The gap between the expectations and the realities of ML is still quite important. Some threats and opportunities are frequently left unaccounted for, execution happens to be more tedious than expected and projects fail to deliver the value that their users and stakeholders awaited.
This research aims at mitigating these issues and improving the outcome of ML projects as well as their predictability. To that end, a breakdown of the structure of ML projects is presented to enable an efficient discussion of each phase of their execution. For each phase, technical and managerial recommendations gathered from the literature, interviews and a survey are provided. In particular, a list of elements to consider while defining the requirements of ML projects and a valuation framework (MLVal) are described.
Results:
- A breakdown of the structure of ML projects
- An aggregation of recommendations for each phase of ML projects
- A list of elements to consider while defining the requirements of ML projects
- MLVal, a valuation framework for supervised ML projects
Summary:
Healthcare takes up a broad range of activities. It can range from the management of acute conditions as a result of a severe health malfunction, or it can be a particular precaution in lifestyle, allowing to minimize future health risks. The later one is becoming more accepted within people, and for this reason, a new market segment is forming. Consumer healthcare is a market where the know-how of traditional pharmaceutical and fast-moving consumer healthcare industries is combined. Numbers of companies are already competing in this industry while providing to consumers with many options of over-the-counter drugs, medical devices, functionalized food, and other personal healthcare management goods. By this work, it will be aimed to deconstruct the external global competitive environment of the consumer healthcare industry from a perspective of one of the leading players in the industry, Procter & Gamble.
Furthermore, the internal strategic analysis of Procter & Gamble personal healthcare business unit will be provided. Finally, the analysis will be reflected in recommendation that could be possibly implemented at Procter & Gamble.
Results:
- Consumer healthcare is a highly segmented industry where regional brands hold 80% of the market.
- P&G has critical weaknesses in geographical diversification and struggling to leverage economy of scale for its products.
- The solution to address its weaknesses is to fill white spaces in the global market by levering its strength in powerful supply chain and a strong portfolio.
Summary:
The role of private investors in renovating Europe’s old infrastructure is a subject National Governments, international institution (World Bank, UN, and IMF) as well as private investors need to set as a major priority. Broadly, countries either have an extensive but old infrastructure that they need to renovate, which is the case of the United States or Europe, or a whole infrastructure network to develop which is the case of emerging countries and regions such as Latin America, South-East Asia, Africa, and the Middle East. The development and maintenance of such infrastructure require massive investments.
Infrastructure investments are traditionally publicly funded. However, public income, generated by tax payments, is used to satisfy the population’s basic needs, the countries’ defense expenses, as well as to cover the running expenses of the public authorities. Public funding for infrastructure development is decreasing and is currently below the actual infrastructure funding needs. To cover this gap, governments must rely on debt financing or tax-increase. These two options are not suited for the current situation of global public accounts: countries tend to reduce their debt/GDP ratio, and tax rates are high enough for them to be not-increased. Finding new funding sources is then essential. Private investors are today the best funding option on the market. Therefore, the goal is to integrate these private investors into funding infrastructure development.
This issue is the core of this research. After defining the problem statement, the first part of the research presents infrastructure as a broad subject. It then describes the state of infrastructure in a predefined geographical area, Western Europe, with a focus on transportation. Financing needs are then studied and defined. The research will show that Italy is in significant need of investments. Therefore, the goal would be to study why Italy is suffering in finding financing needs, and what public authorities can do to push private investors towards this market.
However, before developing this section, Private-Public Partnership (PPP) is detailed and discussed as the latter is the proper tool used when a private entity works together with a public authority on realizing a public asset, which is infrastructure. Most aspects of PPPs are covered. The second part of the research establishes the advantages and disadvantages of PPPs. It then describes the different types of PPPs, the level of involvement of the private and public sector in each PPP type, with a focus on DBFOM type of contracts. The actors and players in a PPP transaction are then presented, followed by a detailed description of the lifecycle of a project together with the bidding process. Project Financing, the financial tool used to finance PPP-type of projects, is presented. A detailed sub-section is dedicated to the concept of risk-sharing between private and public entities as it is a core part of PPPs. Finally, a list of incentives that governments and public authorities should consider to easily attract and better integrate private investors in the infrastructure investment ecosystem is presented.
Finally, this research discusses a case-study. This case-study present the Italian case, with a comparative analysis towards Germany and France, and by using data that will allow to quantify the problem. The findings of this case show that private investors are not rushing towards investing in Italian infrastructure due to the governance and the regulatory aspect of the country.
Italy’s public authorities’ governance is weak, with a high risk of corruption and a low rule of law. The legislative power should play a significant role in detecting the leaks in the legal system that allow local authorities, employees, or major private actors to get involved in corruption schemes. The corruption controlling authority should be enforced, and its activities and scope of application should be reviewed for it to be efficient. The judiciary system can play a major role by enforcing stricter sanctions towards corruption scandals, mainly in the infrastructure sector. Italy is one of the riskier countries, financially speaking, in Western Europe. For authorities to attract non-local private and institutional investors, it should loosen the financial regulations on interested investors, in order to counterbalance this risk-effect.
Apart from these broad reforms, the main issue remains the economic situation of the country.