Digital Waste

IN

DEGREE PROJECT MECHANICAL ENGINEERING, SECOND CYCLE, 30 CREDITS

STOCKHOLM SWEDEN 2019,

Digital Waste

ELIMINATING NON-VALUE ADDING

ACTIVITIES THROUGH DECENTRALIZED APPLICATION DEVELOPMENT

MACHTELD BÖGELS

KTH ROYAL INSTITUTE OF TECHNOLOGY

SCHOOL OF INDUSTRIAL ENGINEERING AND MANAGEMENT

ELIMINATING NON-VALUE ADDING ACTIVITIES THROUGH DECENTRALIZED APPLICATION DEVELOPMENT

Digital Waste

0 1

0 0 0

1 0 1 1 1

1 1

0 1 0 1

1 0 0

1 0 1 1

1 0 0

1 1 1

1 0

Stockholm, Sweden June 2019 MG213X – Degree Project MSc Production Engineering & Management

Department of Production Engineering School of Industrial Engineering & Management (ITM) KTH Royal Institute of Technology Machteld Bögels

Gunilla Franzén Sivard Lasse Wingård

Ron Augustus Author

Academic Supervisor Examiner

Company Supervisor

Abstract

abstract

keywords

In an era where the network of interconnected devices is rapidly expanding, it is difficult for organizations to adapt to the increasingly data-rich and dynamic environment while remaining competitive. Employees experience that much of their time and resources is spent daily on repetitive, inefficient and mundane tasks. Whereas lean manufacturing has manifested itself as a well-known optimization concept, lean information management and the removal of waste is not yet being used to its full potential as its direct value is less visible. A case study was conducted to define which types of non-value adding activities can be identified within information flows and to determine whether decentralized application development can eliminate this digital waste. An internal information flow was modelled, analyzed and optimized by developing customized applications on the Microsoft Power Platform. Based on literature from the field of manufacturing and software development, a framework was developed to categorize digital waste as well as higher order root causes in terms of business strategy and IT infrastructure. While decentralized app development provides the ability to significantly reduce operational digital waste in a simplified manner, it can also enable unnecessary expansion of a common data model and requires application lifecycle management efforts as well as edge security to ensure data compliance and governance. Although limited to one case study, the suggested framework could give insights to organizations that aim to optimize internal workflows by identifying and eliminating digital waste and its root causes.

I en tid där nätverk av sammankopplade enheter expanderar snabbt, är det svårt för organisationer att anpassa sig till den allt mer datoriserade och dynamiska miljön och samtidigt förbli konkurrenskraftiga. Anställda upplever att mycket av deras tid och resurser spenderas på repetitiva, ineffektiva och vardagliga uppgifter. Lean manufacturing har visat sig vara ett välkänt optimeringskoncept, dock har informationshantering och avlägsnande av slöseri inte ännu nått sin fulla potential eftersom dess direkta värde är svårare att se och räkna. En fallstudie genomfördes för att definiera vilka typer av icke-värdeskapande aktiviteter som kan identifieras inom informationsflöden och för att avgöra om decentraliserad applikationsutveckling kan eliminera detta digitala slöseri. Ett internt informationsflöde modellerades, analyserades och optimerades genom att utveckla anpassade applikationer på Microsoft Power Platform. Baserat på litteratur från tillverknings- och mjukvaruutvecklingsområdet utvecklades en ram för att kategorisera digitalt slöseri samt högre grundorsaker när det gäller affärsstrategi och IT- infrastruktur. Medan decentraliserad apputveckling ger möjlighet att avsevärt minska det operativa digitala slöseriet på ett förenklat sätt, så kan det också möjliggöra onödig expansion av en gemensam datamodell och kräver hantering av livscykelanalyser samt kantsäkerhet för att säkerställa datahantering och styrning. Trots begränsad till en fallstudie, så kan det föreslagna ramverket ge insikter till organisationer som syftar till att optimera interna arbetsflöden genom att identifiera och eliminera digitalt slöseri och dess grundläggande orsaker.

samman fattning

nyckelord

2

Contents

Abstract ...1

Abbreviations ... 3

1. Introduction ... 4

1.1 Background ... 4

1.2 Problem statement ... 7

1.3 Research scope ... 7

1.4 Limitations...8

2. Literature Review ... 8

2.1 Information management in organizations ...8

2.2 Digital Waste ...8

2.3 Eliminating waste through automation ... 10

3. Methodology... 11

3.1 Research questions ... 11

3.2 Research methodology... 11

3.3 Software ... 14

4. Analysis, results and proposal ... 14

4.1 Describe ... 14

4.2 Identify ... 17

4.3 Prescribe ... 20

4.4 Develop ... 22

4.5 Implement ... 27

4.6 Evaluate ... 28

5. Discussion ... 31

6. Conclusions ... 33

7. Bibliography ... 35

8. Table of Figures ... 36 9. Appendix ...

9.1 PowerApps code ...

3

Abbreviations

API Application Programming Interface ATU Account Team Unit

BSO Business & Sales Operations CSU Customer Success Unit DFD Data Flow Diagram EOU Enterprise Operating Unit ER Entity-Relationship FY Future Year

MSAccess Revenue reporting/billing engine tool

MSCALC Tool that manages account structure and id’s from MSX and MSSales MSSales Tool that reports revenue

MSX Internal Customer Relations Management (CRM) tool SMB Small-to-Medium Businesses

STU Specialty Technology Unit

4

1. Introduction

1.1 Background

The increased usage of online devices has enabled a global interconnectivity impacting consumers as well as organizations. Companies have found the strong incentive for digital transformation and cloud adoption to optimize internal operations and maintain a competitive position in the market. Data is collected on a continuous basis to obtain strategic and real-time data-driven decision making, imposing significant challenges on maintaining an efficient work environment for employees. Decisions that will impact employees and organizations are preferably made based on accurate and up to date information. In many cases correct information is missing, or a significant amount of time is spent on information gathering and knowledge exchange. It is estimated that on average 59% of managers are missing valuable information daily and in general knowledge workers spend around 20% of their time looking for the right information (Feldman, 2001).

Organizations consist of siloed departments instead of having a shared information environment in which accurate information is always available to the right person or team when required. Although numerous software applications have been developed to support different purposes, typically each department or employee maintains their own working method with respect to file storage and information sharing.

Whereas lean manufacturing has manifested itself as a well-known optimization concept, lean information management and the removal of waste is not yet being used to its full potential as its direct value is less visible. Non-value adding efforts regarding information management can be defined as digital waste, which exists in many forms. Within manufacturing, production line automation has proven itself capable of optimizing production flow and eliminating many non-value adding activities. Similarly, technological solutions that support rapidly changing business needs in an effective and agile manner could potentially provide the same level of optimization for information flows.

Software companies such as Microsoft have grasped the opportunity to provide the ability for employees inside an organization and outside of a typical IT department to rapidly develop applications that can be completely customized to support a specific business process or information flow in a simplified manner. The question that arises is whether these decentralized technological solutions such as automated internal workflows will be able to eliminate non-value adding activities and what would be necessary for such an implementation to be successful.

This research aims to investigate digital waste by analyzing an organizational information flow to determine whether theories from other research fields can be applied to identify and categorize non- value adding activities. The information flow considered within this case-study will then be optimized through automated workflows and decentralized app development, to determine the ability of customized technological solutions to remove digital waste. The software which will be used in the attempt to eliminate digital waste is provided by Microsoft, from wherein this research will be conducted.

5

1.1.1 Microsoft Corporation

The Microsoft Corporation (Redmond, WA, U.S.) is a multinational organization developing and selling computer software, personal computing devices and services. Satya Nadella, the current CEO, initiated a major shift in the company culture, for example through reformulating the global mission statement, namely ‘to empower every person and every organization on the planet to achieve more’.

Over the past years, Microsoft has realized the necessity to move from on-premise software solutions to focusing more on the development of Azure, Microsoft’s cloud platform, on which these existing products can be deployed. There are five main solution areas on which the focus lies within product development at Microsoft: Modern Workplace, Business Applications, Applications & Infrastructure, Data & Artificial Intelligence and Gaming. Microsoft’s ambition is to reinvent productivity and business processes by building and providing an intelligent cloud platform and enabling more personal computing.

As organizations and their operations must adapt to rapidly changing circumstances, the need for customized and agile application development to support specific business processes has increased over the past years. In order to fulfill that need, Microsoft has developed the Power Platform which decentralizes application development and enables employees from different departments and teams to build applications that support customized scenarios, automated workflows that remove repetitive tasks and dashboards that visualize data. These three solution areas are described in more detail in section 3.3.

1.1.2 Microsoft Netherlands

The Dutch subsidiary of the Microsoft Corporation aims to enable organizations in the Netherlands to become an icon for digital transformation worldwide. Over the past decades, companies have realized the need to migrate their activities to a digital and even cloud-based environment to improve the efficiency of their internal and external operations and to remain competitive in a highly dynamic environment. The purpose of products and services that Microsoft has developed is to support and enable organizations to realize their full potential in an effective and agile manner.

Microsoft Netherlands consists of different units which are responsible for varying operations from a business-to-business perspective towards customers. These customers have implemented or are currently deploying Microsoft products including for example the cloud platform (Azure), tools that support business operations (Dynamics) as well as collaboration tools (Office) based on licenses that are set for a certain period. Based on the number of individual licenses (seats) they have deployed, organizations are either considered an enterprise or a small-to-medium sized business (SMB). The Enterprise Operating Unit (EOU) is responsible for the contact and support of managed customers, which are enterprise organizations within the public and commercial sector. For each of these organizations, an account is created to which different Microsoft employees are assigned and involved from three main units: the Account Team Unit (ATU), the Specialty Technology Unit (STU) and the Customer Success Unit (CSU). Every role that is related to an account has a specific orientation and area to focus on as well as their own roadmap, depending on the team and/or unit from which they operate. For SMB customers there are no specific roles assigned to the account, i.e. they are not solely managed, but they are supported batchwise.

6

1.1.3 Current situation

As was described before, the products that Microsoft provides to its customers are typically deployed as a service based on a licensing contract that is established by employees that operate within a sales role. For larger accounts, i.e. based on the number of seats which are licensed through the contract, there are multiple roles assigned within the units that were stated before, i.e. the ATU, STU and CSU.

Whenever a new license is deployed, an invoice is sent to the customer for which an identifier is created (MSSalesID). If that license would be renewed, another invoice would be sent to that same customer and corresponding MSSalesID. Revenue received through invoices lands in MSSales, a tool which is internally used to match revenue to a certain (existing) account. Employees from the Business

& Sales Operations team (BSO) are responsible for managing revenue and ensuring correct account structures.

An important aspect to address within this scenario is that accounts can be subsidiary organizations to other organizations. Revenue that is generated for a subsidiary account also contributes to the total amount of revenue for its parent organization. Customers’ organizational structures are monitored within Microsoft to ensure that revenue is reported in the right location. If an account has a parent organization which is also a customer, the MSSalesID of that parent organization is added to the subsidiary account using an additional identifier, the Top Parent ID (TPID). Over the past years, the number of accounts that are incorrectly positioned in an organizational structure has increased significantly. Employees from the ATU who establish license contracts with these customers are responsible for monitoring their own portfolio in terms of the total revenue that they have generated throughout the year. To reassign accounts to the correct parent organization, a request can be made by an employee from the ATU which is to be approved by an employee from the BSO based on certain conditions.

Currently, there is a Microsoft Excel file that is sent to employees from the ATU, i.e. sellers, in which they can search for accounts that have generated revenue during that month which are located in the wrong place. A lot of time is spent by both sellers who search for misaligned accounts as well as BSO employees who need to approve or reject these requests, as there is no structured workflow in place.

Many emails are sent back and forth about a specific case and sometimes multiple requests are done for one invoice, causing a lot of inefficiency and extra workload. As of now there is no possibility to gain insight in requests from the past or those that are to be reviewed once again the year after.

Overall, a lot of time is spent on mundane and inefficient tasks such as gathering, managing and distributing files related to these account revenue requests. It is an unstructured but reoccurring process which can potentially be optimized if the right approach is used.

7

1.2 Problem statement 1.2.1 Definition

The problem that is defined within the scope of this research is that employees from the BSO and ATU teams experience that a lot of their time is spent on non-value adding activities regarding account revenue requests. It is an unstructured process that consists of a lot of inefficient, repetitive and mundane activities which negatively impacts daily operations for many employees.

1.2.2 Stakeholders

From an operational perspective, the main stakeholders within this scenario are BSO employees who are responsible for reviewing requests and ATU employees, as they are responsible for their own customer portfolio and corresponding revenue. On a more strategic level, Microsoft as a whole is considered another stakeholder in this scenario as the organization aims for high employee efficiency as well as having up to date organizational structures in terms of revenue reporting. Additionally, investigating whether Microsoft’s customized application development platform provides the ability to remove non-value adding activities internally could be beneficial to understand its impact and how to support customers in similar situations.

1.3 Research scope

To improve the existing level of internal collaboration and knowledge exchange, it is necessary to critically reflect on internal information management efforts. Analyzing data flow from collection to strategic decision making could gain the necessary insights in how data travels through organizations and whether this data and its related activities add value to the overall process or not.

The aim of this research is to identify non-value adding activities through information flow modelling in order to gain insights that support a potential categorization of digital waste and to optimize and eliminate activities that are currently counteracting a collaborative work environment. It is expected that the information model that is to be developed, assuming that it represents the actual functionality of the real information system, can provide insights in the amount of wasteful activities and potential root causes that exist within the described scenario.

Analyzing the similarity between waste that is found within production lines and information flows can be useful to determine whether optimization strategies such as lean manufacturing could be applied in information management scenarios to eliminate wasteful processes. An optimized model will be designed and developed in which information is exchanged in a structured and more efficient manner using customized technical solutions. An analysis will be made to determine whether this optimized information system contains less non-value adding activities and which prerequisites hold for such an implementation to be successful.

8

1.4 Limitations

This research aims identify which types of digital waste exist within a certain information exchange scenario at Microsoft Netherlands to analyze the impact of a technical (automated) solution on the level of digital waste. The research is narrowed down to one specific scenario which can be modelled and for which an end-to-end solution can be developed using Microsoft products, inherently evaluating the ability of these tools to solve internal situations in which digital waste is found. The usage of these software applications limits the research in the sense that only these tools are evaluated for their ability to support automated workflows and elimination of digital waste. A more detailed description of the Microsoft tools that are used in this research is given in section 3.3.

2. Literature Review

2.1 Information management in organizations

As organizations are currently in their digital transformation journey, the amount of data that is generated and collected daily increases at a significant rate. Long-term business sustainability depends on the ability to acquire knowledge throughout an organization and which can promote the development of better products and production processes (Lodgaard, 2018). Appropriate information management strategies can provide the necessary guidelines for optimizing internal knowledge exchange. Information modelling can be used to gain insights current informational ecosystems and can be a foundation for software development. Different information models based on how they can be used to describe information exchange processes within organizations (Durugbo, Tiwari, & Alcock, 2013). Where information modelling efforts were previously aimed at improving its efficiency, there is now an increasing interest in evaluating information management efforts based on adaptability and flexibility. Considering that there are numerous standards for information modelling it depends on the specific requirements for the outcome of the modelling efforts to determine which standard is most suitable. In order to identify data latency, the Data Flow Diagram (DFD) is considered most suitable information model to use (Hoitash, 2006). Within the DFD, information is modelled through processes and data stores, enabling the identification of manual data entry as well as the different data formats and software applications that are used. Developing and understanding the flow of data in and through the system given the complexity of work processes is a challenging task (Murray, 2003).

2.2 Digital Waste

Toyota developed the concept of lean manufacturing which has been widely applied throughout the production industry and has been introduced in other areas as well. Within this theory seven categories of waste are identified, which are all non-value adding activities related to transport, inventory, motion, waiting, overproduction, over-processing and defects. Additionally, an eighth category is identified which describes the waste of not using the talent and capabilities of humans, i.e. the employees within your organization (Liker, 2004).

9

The principles of lean thinking, the removal of waste and the pursuit of perfection can be applied to any system where products flow to meet the demand of a customer, user or its consumer (another system). More specifically, it can be applied to information management since information typically flows through an organization and related efforts are aimed to add value to the product, e.g. when data is generated to lead to or support a certain decision that is to be made (Hicks B. , 2007).

Fundamental to successful application of lean is the identification of value, understanding of flow and characterization of waste. Within information management, identifying value as well as characterizing waste is a complex task since it is less tangible and highly subjective. This becomes especially clear when compared to a manufacturing environment where value and waste, due to its visibility, can be measured in key performance indicators, giving direct insights in flow efficiency. The potential barrier of understanding value and waste is important to consider when modelling information flow and developing possible improvements, as its effects are difficult to measure objectively.

Figure 1 conceptualizes a pyramid of knowledge: the raw data stored in a database will add value towards the decision that is to be made only if the right information is presented in the right format to the right people at the right time. It must be structured and presented as digestible information such that a human can interpret it and knowledge is created. Over time, as knowledge is accumulated and combined with experience and judgement wisdom is developed. (Bell, 2006)

Structuring and filtering raw data such that value-adding information can be presented to the right person can be done by implementing and adopting the right technology, which can be challenging within multidisciplinary organizations.

Evaluating information management issues within ten small and medium sized enterprises has led to the identification of 18 core issues that occurred among these organizations (Hicks B. C., 2006). After reevaluating it was determined that these issues were caused by four fundamental waste categories (Hicks B. , 2007):

• Failure demand: the resources and activities necessary to overcome a lack of information.

FIGURE 1– PYRAMID OF KNOWLEDGE (BELL,2006)

10

• Flow demand: the time and resources spent trying to identify the information elements that need to flow

• Flow excess: the time and resources necessary to overcome excessive information, i.e.

information overload

• Flawed flow: the resources and activities necessary to correct or verify information as well as the unnecessary or inappropriate activities that result from its use.

Additionally, these waste categories were mapped directly onto the types of waste that were defined for the Toyota Production System, suggesting that Failure demand, Flow demand, Flow excess and Flawed flow are similar to over-processing, waiting, overproduction and defects. The remaining waste categories, namely transport, inventory, motion as well as not using people to their full potential are in this study not considered as digital waste.

Others argue that digital waste is to be defined beyond non-value adding activities by categorizing it as having either a passive or active nature. Passive digital waste occurs when digital opportunities are missing to unlock the power of (existing) data. Active waste on the other hand, results from a data- rich environment that lacks the appropriate information management approach to derive the right information to be provided at the right time to the right person, machine or information system for decision-making (Romero, Gaiardelli, Powell, Wuest, & Thürer, 2018). Simultaneously, digital waste can also be described more literally within four categories (unintentional data, used data, degraded data and unwanted data) to which a waste elimination approach could be applied similar to how physical waste (from households for example) is handled in daily life (Hasan & Burns, 2013). It can be concluded that there is no significant consensus on the definition of digital waste, due to the fact that it depends highly on the situation and environment to which it is applied.

2.3 Eliminating waste through automation

In terms of waste elimination strategies, automation arises as an optimization strategy for production lines which has been widely applied across the industry and is being developed constantly. Automation can be described as the ‘automatically controlled operation of an apparatus, process or system by mechanical or electronic devices that replace human labor’. One distinction that must be made is whether the process that is to be automated consists of continuous or discrete events, as it results in either process automation (continuous) or factory automation (discrete). Another categorization that can be made is whether automation is considered hard, i.e. when an industrial robot is programmed to perform one specific task, or soft, i.e. when more flexibility is required through the ability to perform different tasks (Wilson, 2015).

Automating manufacturing systems improves productivity and the overall efficiency of a production line as well as a significant increase in quality due to the specific tolerances that can be applied to automated assembly systems. Moreover, it replaces repetitive and mundane tasks previously performed by humans. Successfully developing, implementing and maintaining automated solutions is critical to optimizing a production line, as it should not lead to jeopardizing critical operations or processes. A significant challenge during development is translating business process into system logic which is then supported by an automated workflow system (Murray, 2003).

11

An important aspect to consider regarding is the involvement of people that will be using the solution in the end, e.g. production employees. They understand the difficulties and variability of the system and will therefore be able to provide useful insights during the development and implementation phase (Wilson, 2015). Another crucial aspect of successful adoption is organizational support and employing a cross-functional implementation team. Additionally, understanding the impact of workflow automation on the organization is important to consider as well as the human interaction and participation intrinsic to such solutions (Murray, 2003).

3. Methodology

3.1 Research questions

Based on previous research it can be stated that the definition of value within information management is inconclusive due to its invisible nature and lack of measurability in terms of monetary value. This leads to the suggestion that the first research question is to be formulated as: how to define value within the concept of information management?

Furthermore, if there would be a common agreement on the definition of value, would there be a possibility to find significant similarities between a production line and an information flow and more specifically: which types of digital waste can be found within information flows?

Identifying and classifying potential types of digital waste are important steps in optimizing information processes such as the scenario described within the scope of this research. The Power Platform, developed by Microsoft to simplify decentralized application design and enable end-users to develop their own tools that support their specific operational business needs, can then be used to investigate the final research question: can decentralized application development be used to eliminate digital waste?

3.2 Research methodology

Since this research is aimed at modelling a complex system by collecting information based on interviews with employees inside the organization, different views on the systems and/or organization can be expected based on experience and perspectives. Therefore, a Soft Systems Methodology (SSM) approach (Checkland & Poulter, 2006) is used within the scope of this research since there might not be one unified solution or suggestion to the problem that was stated before. The methodology is used to identify and describe a situation to create (conceptual) models that describe the behavior of the actual system, especially within complex systems where variables are unknown and the system can be viewed from many perspectives. Capturing the overall functionality and behavior of an information flow model in terms of people, information and the technology used is a complex task. Therefore, the research is aimed towards modelling one specific scenario to be able to identify challenging or wasteful areas that most likely exist.

12

The methodology that was selected to conduct this research is an interpretation of the engineering design process (ITEA, 2007), which is a systematic problem-solving strategy that aims to develop a solution that meets predefined requirements or demands. Figure 2 shows the research methodology that was developed for this research which is an interpretation of the engineering design process.

In this scenario, the describe, identify and prescribe steps combined lead to a set of requirements which is used to develop, test and potentially redesign a solution before implementing it into the real scenario and evaluating its impact. A more detailed description of each phase is given in the next sections.

3.2.1 Describe

The first step is to describe the current status of internal collaboration and how data is handled by creating an information flow model according to the standard of a Data Flow Diagram (DFD) which consists of processes, data stores and conditional statements. These conditional statements will provide the ability to identify decision points and introduce logic that determine whether, in this specific scenario, requests are to be approved or rejected resulting in different subsequent flows. With a DFD model it will also be possible to evaluate the flow from different perspectives to identify which role is responsible for which part of the information exchange and how this potentially could be optimized.

3.2.2 Identify

After modelling the information flow, it is necessary to define value within the concept of information management which will be used to determine which activities are considered value-adding and which are not. Given an established definition of informational value, different types of waste can be identified as well as to what extent they exist within the flow, necessary to answer the first two research

FIGURE 2– RESEARCH METHODOLOGY

13

questions. Additionally, suggestions for potential root causes can be made, enabling stakeholders within this research to critically reflect on their information flow in other scenarios as well.

3.2.3 Prescribe

After that, the aim is to optimize the system by eliminating non-value adding activities and developing an information model that prescribes how information can be exchanged in a more efficient manner using technology. This will be done by developing an optimized DFD-model including only necessary and value-adding activities including required data stores. These data stores are then incorporated into an Entity-Relationship (ER) model that shows how various data sources are linked to each other.

Resulting from this phase is a set of requirements necessary to cover the overall functionality of the optimized scenario which is used to design and validate the solution that is to be developed.

3.2.4 Develop

The next step is to use the ER-model to develop a solution using different Microsoft products from the Power Platform, i.e. PowerApps, Flow and PowerBI. The specific functionality of these products will be described in section 3.3. The aim is to provide an end-to-end solution that automates the non- value adding activities which were identified before. The functionality of the developed solution is tested by applying a use case scenario. The prescribe, develop and test phase are part of a sub- iterative cycle that is aimed to optimize the suggested solution. It is expected that during development the solution will be tested and feedback will provided such that changes to the ER-model will be conducted, developed and repetitively tested in an agile manner until the solution meets all requirements.

3.2.5 Implement

After the solution is validated through a proof-of-concept, it can be implemented to replace the real workflow. Successful adoption of the solution requires a thorough strategy that incorporates every aspect of the implementation such as modifications to fit the actual situation, i.e. connecting specific roles to tasks. Other elements include planning the launch to avoid corrupting daily operations but also ensuring the adoption by employees through clear instructions and managing changes appropriately.

3.2.6 Evaluate

The aim of this phase is to evaluate insights that were gained during the different steps that were conducted within this research, to determine the validity of potential types of non-value adding activities that were found and to establish which prerequisites hold for an implementation of Power Platform solutions to successfully eliminate digital waste, contributing to answering the second and third research question. Another important aspect to consider is how these insights can be used to provide feedback to the prescribe phase, enabling potential improvements on this optimization strategy when applied in other scenarios.

14

3.3 Software

Microsoft has developed a solution called the Power Platform which consists of three products, i.e.

PowerApps, Flow and PowerBI that have separate functions but can be combined to provide a customized operating platform.

PowerApps is a cloud-based application that enables users to develop a personalized application that consists of built-in functionalities, simplifying the application development process while focusing more on the customization aspect. It has a built-in common data model which provides a database in which entities are stored and data can be referred to and edited from multiple locations.

Additionally, it can establish connections to read and write data from over two hundred external data sources such as Microsoft OneDrive, SQL servers as well as online services such as social media sites and weather applications.

Microsoft Flow is a tool that can be used to automate repetitive tasks and business processes by connecting different applications and databases to each other through connections and APIs based on predefined workflows or triggers.

PowerBI is a visualization platform that can give end-users actionable insights through customized dashboards based on data collected through different sources and connectors such as SQL servers, PowerApps and Flow.

4. Analysis, results and proposal

4.1 Describe

4.1.1 Account Revenue

When revenue is generated on either a managed or unmanaged account, an entry is made within MSSales, a tool that manage sales and account revenue globally. If revenue is generated for a new account, a unique identifier is created, i.e. the MSSalesID. It could also occur that revenue is generated for an unmanaged (small to medium) account which is a subsidiary of a managed (enterprise) account.

FIGURE 3– OVERVIEW OF APPLICATIONS AND DATA MODELS IN THE POWER PLATFORM

15

That information is stored in the MSCALC tool, which manages and combines information from both Customer Relationship Management (CRM) software as well as MSSales data. In MSCALC, unmanaged subsidiary accounts are assigned as a child organization to a managed parent organization. A significant amount of managed accounts (parents) have multiple subsidiaries related to them as child accounts. Whenever revenue is generated on a subsidiary account, matching occurs to ensure that the correct MSSalesID of the parent organization is added to that specific data entry as a Top Parent ID (TPID). In those situations, the revenue of the unmanaged account is recorded under the managed account, as it contributes to the overall revenue that was generated for the Enterprise Operating Unit and partially determines the account budget that is to be defined for next year.

4.1.2 Requests

According to employees from the BSO team, existing accounts are often not recognized when revenue is generated as well as accounts not being matched to the right parent account. In those situations, a novel MSSalesID is generated and the account becomes its own parent, i.e. the Top Parent ID (TPID) is the same as the MSSalesID and it will be considered an independent SMB-account. On a monthly basis, employees from the Sales Operations team send an Excel file to ATU sellers consisting of all novel SMB-accounts that have been generated for that month, of which some are potentially wrongly considered to be a SMB. Within that list, sellers can search for accounts that should have been assigned or parented to one of their managed accounts. If that is the case, they can submit a parent request by adding a name of the desired parent to the entry within that Excel file. An employee from the BSO team then approves or rejects this request. One of the conditions for approving parent request is that the account (or MSSalesID) should have been created less than two months ago. When this period has passed, employees can still submit a request, but it is less likely to be approved. Other requirements for approval include that the subsidiary organization should be owned by the desired parent organization for more than 51%, and that no previous revenue was generated for the subsidiary organization in the past three years. For each request, employees from the BSO have to search for information that either supports or refutes these conditions leading to a decision to either approve the request, approve it to be included into next year reports or reject it. In some situations, organizations no longer exist as a subsidiary to another organization, i.e. they have become independent and therefore should be assigned to themselves as a parent. There are scenarios in which organizations have merged but the new (merged) organization is still a customer for which revenue could be generated. It is then necessary to submit a merge request to combine both accounts without losing existing information about previously recorded revenue. One organization then becomes the victim organization and the other the survivor, the latter of which the MSSalesID remains.

Additionally, sellers often do not prioritize the search for missing child accounts until the end of the fiscal year when they realize that some of their revenue is not recorded. In many situations the two- month period has then passed so the account remains unassigned. Over the years, these factors have contributed to an incrementation of the number of unassigned accounts until the point where it now consists of around 700.000 entries.

16 4.1.3 Data Flow Diagram

Figure 4 shows the Data Flow Diagram for the scenario that is described. The processes depicted in the blue shaded area within the DFD are the most repetitive and time-consuming and therefore define the scope of this project for information flow optimization. The black arrows depict the process flow whereas the grey arrows describe the flow of information. Processes 1 until 4 describe the steps that need to be taken to create a database with all necessary information about the account, e.g. name, address, account manager as well as its current parent-child structure and recommended parent. In processes 5 and 6 an Excel file is created with SMB-accounts that have landed that month. The actual monthly revenue (visible in the cloud in Figure 4), is reported in MSSales and added to the database in process 7. Prospected revenue for the coming year is reported in MSAccess, an additional reporting tool. These values are visible for the BSO team through an Excel plugin in which they can run queries to export the specific revenue that was created within SMB-accounts and add that to the database in processes 8 and 9. In process 10, the recommended parent from the SQL database is added to the Excel file. During process 11, the BSO team then sends this file to the sellers by email including empty columns in which e.g. a new parent can be stated as part of the parent request. They can search for missing accounts based on for example their name, MSSalesID or revenue that was generated (process 12). When a suspected missing account is found, the suggested new parent is written down in the additional columns, and the request is sent back to the BSO team by email (process 13). In process 14, the BSO team reviews the request based on certain predefined conditions which can lead to different outcomes. In some specific situations, there exists doubt whether a request should be approved or rejected, so it is reassigned to the SMB-team being responsible for all small-to-medium businesses (unmanaged) accounts (process 15c). They can then decide whether it is allowed that that specific

FIGURE 4– DATA FLOW DIAGRAM OF CURRENT SITUATION

17

revenue contributes to the overall EOU revenue or if it should remain at SMB (process 16b). The ATU seller is informed on the outcome of the request(s) through manual emails (process 16a). In process 17 and 18, the requests that were approved are then updated by copying the correct new MSSalesID’s into MSCALC, which is done by someone from the BSO team. The last step in this scenario is for another member of the BSO team to approve the changes that were made in MSCALC (process 19).

4.2 Identify

4.2.1 Value definition

To be able to identify different types of waste that exist in the Data Flow Diagram is it necessary to define value in the context of information flows. As described by (Bell, 2006) and depicted in Figure 1, information is data which is stored and structured until it is interpreted by someone enabling it to become knowledge. Over time, this knowledge gets transferred into wisdom based on experience and insights. Within this transformation, data could be considered the raw material of a production line to which more value is added until it becomes a finished product, i.e. wisdom. On the contrary, as information is less tangible and more subjective in its nature, a less quantifiable definition would be more appropriate. Moreover, as information travels throughout different departments and levels within an organization, the perceived value of a certain set of data or acquired knowledge changes.

The concept of informational value can be viewed from many different perspectives as well as the level on which it is determined, i.e. whether it is on a local, regional or global level as well as an operational or more strategical level. For example, data that seems unnecessary and therefore invaluable within an operational setting could be useful on a long-term on a higher level to gain predictive insights and determine data-driven strategies. The scenario that is depicted within this research occurs on a more local and operational level, in which data as well as information typically only provides value after it has been interpreted, e.g. by a human being or machine learning models.

In order to achieve a consensus on value definition within the scope of this project, it is therefore suggested that activity that is described in the DFD which does not require knowledge or wisdom, is a non-value adding activity. Based on this definition an evaluation is made on the DFD processes considered within the scope of this project. The activities which do not add value are shown as the red colored processes in Figure 5 implying that around 58% of the tasks can be considered digital waste.

The activities which are depicted in white require employees from either the BSO, ATU or SMB team to interpret and evaluate the information that is presented to them during this process, e.g. reviewing or approving a request.

FIGURE 5– NON-VALUE ADDING ACTIVITIES IN THE DATA FLOW DIAGRAM

18 4.2.2 Types of waste

The activities that are depicted in Figure 5 are highly repetitive and require a significant amount of time and resources spent by the BSO team, as it consists of verifying, checking and correcting a lot of information. Processes 11, 13, 15c, 16b and 16c all include communicating with employees through emails regarding the potential approval of requests based on information that usually requires additional searching. There is no historic data on previously submitted requests which means that in many cases emails, searches and requests are done multiple times, creating additional effort and time spent by all employees involved. For requests that cannot be approved or rejected based on given information, it is necessary to communicate with the SMB unit to perform additional reviewing. Based on the identification of digital waste by (Hicks B. , 2007), especially the wasteful efforts and resources related to data verification, correction and duplication frequently occurs within this scenario. Many accounts are unassigned or assigned to the wrong parent and requests are done multiple times due to a lack of monitoring. The monthly repetition of sending Excel files with new revenue that was generated for SMB accounts leads to a lot of time spent on file transfer and version management. This frustrates the involved employees as they experience that these activities are not adding any value to their operations and have a negative impact on their daily routine.

For some processes it seems that the technology used is lacking, i.e. the Excel files with around 700.000 entries which are not filtering correctly is most likely caused by the software not functioning according to these requirements. The efforts and resources spent on data duplication and verification are not only caused by inappropriate technological solutions but are also due to a lack of information management, i.e. the process that determines the setup of this scenario and how information flows throughout the organization. The people, process, technology framework has been widely applied in organizations to improve software development and implementation (Chen & Popovich, 2003). It implies that each aspect within this framework is crucial for any application to be successful, as the technology itself should be functioning but can simultaneously be useless if it not adopted by the people or if it supports improperly designed processes in the first place. In an era where the significance of data and data-driven insights for strategic decision is constantly increasing, it can be suggested to expand this framework into a more current interpretation of what is necessary for successful digital transformation.

Figure 6 represents an interpretation of the people, process and technology framework with the addition of categorizations of digital waste as depicted in previous literature. As data affects and is affected by all three aspects of the framework, it is considered to exist in between the people, process and technology elements. The definition of active and passive digital waste as described by (Romero, Gaiardelli, Powell, Wuest, & Thürer, 2018) provides a useful categorization to be applied to the framework suggesting that a lack of technology can cause passive waste and that insufficient information management on the process side can generate active waste.

19

Additionally, the various non-value adding activities that were found within the scope of this research can be defined according to the classic categorization of waste as defined by (Liker, 2004) for the Toyota Production System. The traditional eight types of waste are also applied to the framework presented in Figure 6 and a more detailed interpretation that was defined within the analyzed information flow is presented in Table 1.

Active Waste All time and resources spent on:

Motion verifying data, i.e. through interpretation by a certain person/tool Overproduction duplicating files and handling excess data

Waiting finding and overcoming a lack of data Defects correcting faulty data

Passive Waste All time and resources spent on:

Transport moving data to the right location, e.g. file handling Inventory migrating and handling legacy data

Over-processing manual data entry

Unused potential Workarounds, e.g. Shadow IT and unused data TABLE 1- CATEGORIZATION OF DIGITAL WASTE

Within this scenario the non-value adding activities that are considered motion, overproduction, waiting and defects can be attributed to inadequate processes or information management efforts and are therefore categorized as active waste. Wasteful tasks that are considered transport, inventory, overproduction, over-processing and unused potential in this information flow are caused mainly by a lack of sufficient technological solutions or integrations, implying that they are defined as passive waste.

FIGURE 6– FRAMEWORK OF FACTORS CAUSING DIGITAL WASTE

20

The types of waste that are highly occurrent in this scenario are transport, motion and defect related activities as most time is repetitively spent on file handling, verification and correcting faulty data.

Additionally, the unused potential of data that is collected through requests which could have useful insights is a significant waste as well as the amount of time spent on filtering and searching within the overproduced amount of data entries within the current account database. Beyond the scope of this scenario, legacy data is suggested as an inventory type of waste as it consists of large quantities of data which typically require significant resources for it to be migrated and accessed and by that complicate overall data transferring processes. Shadow IT was added as an additional form of passive digital waste which can be found when a novel software application is implemented without successful adoption, i.e. the technology is not used to its full potential and therefore creates inefficient and insecure workarounds.

In typical production flows, these types of waste can be eliminated by improving either the process itself or the underlying technology through for example automated solutions that support these processes. When it comes to information flows, it seems that the people involved in handling the information play an important role in the level of waste that can be found. When more people are involved in an information flow, logically a higher level of subjectivity can be expected. People evaluate and interpret information differently which is useful in many situations but can also add a level of complexity when it is not necessarily required. The information management processes are usually defined by people as well, with the right technology implemented to support these processes. In order to improve both the people and the process elements, i.e. both active and passive waste within the framework, it is necessary to critically reflect on the technology that is used and whether it could be optimized in order to remove the resulting digital waste.

4.3 Prescribe

4.3.1 Functional Requirements

To significantly optimize the existing information flow, it is necessary to develop a robust technological solution which automates the non-value adding tasks. The main group that is targeted in this scenario are the sellers (account managers) to whom currently the monthly dataset of newly created SMB- accounts is sent. Instead of receiving a monthly update, the possibility should exist to search in a database which is automatically updated and consists of all potentially misaligned accounts. Two search scenarios are considered to distinguish between managed accounts and their related child accounts as well as the unmanaged child accounts with corresponding parent accounts. Additional information which is necessary for the sellers includes the currently assigned account manager of the (managed) account, the revenue billed from the previous three years as well as the current or expected revenue for this year.

A recommendation for the account which is most likely to be the right parent account for a certain account should be included to simplify the matching process. They are then to submit parent or merge requests which are automatically available to employees from the BSO team consisting of all the required information for them to approve the request, reject it or approve it for the future year (FY).

21

Requests which are approved should then be stored in the right format directly for the parent-child structure of the accounts to be updated automatically in the MSCALC tool. Additionally, whenever a request is updated with information from the approver(s), the seller should be informed directly with the approval status and corresponding comments.

4.3.2 Data Insights

The information regarding these requests, i.e. whether they have been approved or rejected including commentary provided by the BSO team should be stored such that requests cannot be duplicated and no communication is necessary to discuss how requests were handled for what reason.

Additionally, requests that are approved for the future year should be stored as well to ensure that no information is lost in the process and necessary insights can be gained through visualization.

4.3.3 Optimized Data Flow Diagram

The optimized data flow diagram that incorporates the aspects described in section 4.3.1 given in Figure 7. The activities that included communication through email are removed and will be done automatically. In order to have one common data model, the suggestion is to have the databases which will be used to store all the current account and create parent and merge requests located on the Microsoft SQL Server such that this information can be accessed directly by the BSO team.

As was described before, the technology that will be used to develop this solution is based on the Microsoft Power Platform, i.e. PowerApps, PowerBI and Flow. PowerApps has numerous existing APIs through which different applications can be integrated. Unfortunately, there is currently no integration with the MSCALC tool, which implies that updating the parent-child account structures will still be done manually using an Excel file in the right format. Nevertheless, the suggested DFD still shows that the total number of manual activities is reduced by half and that non-value adding activities have been reduced from 58 to 28 percent of the total number of tasks. Simultaneously, this does not necessarily imply that the amount of total time and effort spent in this scenario will reduce in a similar manner, as they are two separate use-cases which are difficult to perform a quantitative comparison on before implementation.

FIGURE 7- OPTIMIZED DFD

22 4.3.4 Entity-Relationship model

The suggested data flow diagram as described in Figure 7 consists of four data stores, i.e. entities which are necessary to conduct the processes in the optimized information flow. Figure 8 shows the Entity Relationship model which is used to visualize the data model supporting the data flow diagram.

It includes all the existing entities necessary to retrieve information from MSCALC, the SQL server, MSAccess and MSSales to generate an up to date account database as well as additional information about internal users of the application. Additionally, the four entities required to support the DFD as suggested in figure 7 are highlighted in the blue are. These consists of three SQL databases: one consisting of the current accounts, one to create parent requests and one for merge requests. The fourth additional data store is used to retrieve user information through an existing API between PowerApps and Office365. For simplicity only the primary key which relates each entity to one another is included in the ER-model, namely either the OrgID or MSSalesID, both referring to one instance of an account.

4.4 Develop

To support all requirements defined in the previous section, two applications are developed within Microsoft PowerApps with customized functionalities for which the developed code can be found in appendix 1, ensuring the possibility to recreate the designed solutions. All images depicted in this section are anonymized to ensure GDPR-compliancy regarding customer data.

FIGURE 8– ENTITY-RELATIONSHIP MODEL

23 4.4.1 CALC_Cleanup

The first application is developed for account managers, forming a target group of around 40 to 50 employees. As was stated before, two search scenarios are suggested from an account manager, i.e.

through managed and unmanaged accounts. Figure 9 shows the designed screen in the CALC_Cleanup application including a gallery on the left side showing all managed accounts including their segment (public or commercial) their subsegment (specific industry) as well as the responsible account manager. This account database is updated daily, showing the most current version of the parent-child structure as it is stored in MSCALC.

FIGURE 9- MANAGED ACCOUNT SCREEN IN CALC_CLEANUP APPLICATION (ANONYMIZED)

There are multiple filtering options designed to simplify the search for any specific account, such as the option to only show the accounts managed by the seller who is logged in. When a managed account is selected, all related child organizations are shown in the gallery in the middle of the screen.

When a specific child organization is selected, there are two options for the user to choose from: to create either a parent or merge request. When a new parent is required, the user can either find a parent in the dropdown list or select the unparent checkbox, implying that the current child organization becomes an independent account and the Top Parent ID will be the same as the MSSales ID. This scenario will be used mostly to remove child organizations which are mistakenly assigned to the wrong parent organization or for organizations that have become independent.

24

Similarly, Figure 10 represents a separate screen which is designed to show all the unmanaged accounts, including all the SMB-accounts which are mistakenly independent but instead should potentially be assigned to a managed account.

FIGURE 10- UNMANAGED ACCOUNT SCREEN IN CALC_CLEANUP APPLICATION (ANONYMIZED)

As was stated before, there are some conditions which predefine whether accounts are eligible to be parented by a seller or not, such as whether the account was created within the last two months, if there was no revenue billed the last three years and if for example the child organization is owned by its parent organization for more than 51%. If either one of these conditions is not met, the Parenting Allowed property will become ‘No’ which informs the seller beforehand that it is unlikely for the request to be approved. It does not prevent the request from being submitted as some requests can also be approved for next year. The seller is obligated to add comments to validate the request, otherwise the submit button will provide an error message. Additionally, when an organization is selected, a search query is executed in the parent request database to determine whether a request for that account has been done previously including the new parent that was requested and what the status for that specific request is. In that way duplication of requests is prevented as the same request cannot be submitted twice.

25

Figure 11 shows a third screen which is designed for the CALC_Cleanup application, showing the user’s requests and their status including comments made by the approver.

FIGURE 11- REQUEST SCREEN IN CALC_CLEANUP APPLICATION (ANONYMIZED)

26 4.4.2 CALC_Approvals

The second application that is developed to support the prescribed scenario is the CALC_Approvals application which functions as a back-office application for BSO employees. Figure 12 shows a screen with pending requests which are created in the CALC_Cleanup tool and can be approved directly, approved for next year or rejected along with explanatory comments made by the approver.

Approved requests can be filtered and when the export button is pressed, all necessary information will be patched to an Excel file consisting of the exact format used for importing in the MSCALC tool.

In that way, the number of manual activities is minimized as much as possible.

FIGURE 12– OVERVIEW REQUESTS IN CALC_APPROVALS APPLICATION (ANONYMIZED) 4.4.3 Challenges during development

The main challenge during design and development of these solutions was to establish connections between the SQL server and Power Platform using an on-premise gateway. The properties within the entities that were interconnected through this gateway consisted of different formats and naming structures which caused some inefficiency due to a lack of standardization. Outside of the scope of this particular project, the decision was made to move from an on-premise SQL server to a cloud- based server, implying that the gateway will have to be established once again.

Applications within PowerApps are always developed in a specific environment, either for trial, development or production purposes with corresponding functionalities. The production environment is most suitable for applications which are used by multiple employees that need to be scalable and secure for actual usage. If a production environment is to be used internally at Microsoft, a specific request must be made to monitor which employees are building which tools for what purpose. This turned out to be another hurdle in the development process as it required more time than anticipated.

Another challenging aspect is that an on-premise SQL gateway is not supported in a production environment, implying that the cloud gateway has to be established before further testing and development in the production environment could be realized.

27 4.4.3 Testing

Development of technology to support a certain process for people requires the involvement of the stakeholders which are to critically reflect on the process that they are using. It gives insight to the end user in what data model is supporting their daily operations. This resulted in a series of internal meetings with the BSO team to state the requirements which were then translated into development functionalities in the tool. These functionalities were then presented in following meetings in which feedback was provided in a cyclic and agile manner until all requirements were met and no additional improvements were necessary.

4.5 Implement

4.5.1 Scalability and adoption

As was mentioned before, the target group to use the applications are account managers and employees from the BSO team. Some important aspects to consider during the implementation phase are the deployment, scalability and permissions for the designed solutions. Additionally, to ensure that the applications will be used, an appropriate adoption strategy is to be determined as well as a plan for managing the applications and incorporating potential changes and updates after implementation.

For the applications to be deployed amongst multiple teams and departments, it is necessary to establish a secure production environment on the Power Platform in which the applications can run.

The applications are aimed to be launched at the start of the new fiscal year, but for testing purposes it is necessary to give read/write rights within the production environment to different users to simulate a real-world scenario.

For a novel tool to be successfully adopted, it can be useful to establish early adopters within the organization, i.e. stakeholders in the ecosystem who can play an important role in engaging and motivating other employees to use the tool. These people could be for example in a managing role who motivate their own team members, or employees within the team itself who get a first glimpse at the tool and then use their experience to get other team members on board. An introductory session with a presentation and/or workshop can provide a simple explanation on the functionality of the application and the scenarios that it involves.

4.5.2 Application management

An important aspect is to investigate the actual usage of the tool and to determine which teams use the tool more than others and which causes can be identified. It is crucial to share knowledge and experience about the applications in terms of how they are structured and how potential changes and updates can be made easily. This is necessary to ensure that this knowledge does not remain siloed but that instead multiple admins can edit the application and a more agile software developing environment is established.

28

4.6 Evaluate

The case study that was examined within this research provided an example of how digital waste can be found in day-to-day operations and how it makes employees less efficient in terms of time and effort spend on non-value adding activities. The tool that was developed to eliminate these activities will be implemented beyond the timeframe of this project, so a thorough evaluation of its impact cannot be made at this point. Nevertheless, more insights on the identification of potential root causes of digital waste can be investigated further.

Throughout the whole organization, a lot of information is collected and stored about accounts in different manners. For example, information that is collected regarding potential sales opportunities ends up in the Customer Relationship Management (CRM) tool, whereas information regarding revenue is stored in the MSSales tool. In both tools, a specific identifier is created although they refer to the same account. The MSCALC tool, which is used to form a bridge between both tools, can be considered a master data management effort ensuring that all information related to an account is stored in one place to create a single point of reference for different departments, employees as well as applications. The inaccurate alignment of new revenue onto existing accounts is mainly caused by a variety of billing systems in which revenue can be reported, depending on the type of product for which a license is deployed. This mismatching within MSSales occurs on a global level in the organization and generates a bulk of erroneous account structures in the MSCALC tool. The tool that was developed within this research may optimize daily operations for the subsidiary in the Netherlands, but the technology and processes that originate on a global or strategic level are within this scenario the underlying causes for this type of operational waste.

This leads to the suggestion that, from an operational perspective, the framework for digital waste should include a hierarchical aspect to incorporate the level on which the root causes of waste can be identified, whether it is operational, tactical or strategical as well as local, regional or global. The suggested framework is presented in Figure 13, in which the lower, operational level in the pyramid consists of the types of waste as presented in Figure 6. These are the types of waste that are found in daily and repetitive information flows such as the one investigated in this project. Solutions such as the Power Platform can be used to rapidly automate many of these tasks as they are able to support specific and customized business needs or processes with a short and adaptive lifecycle. These solutions are also used mainly on a local level such as the CALC_Cleanup application that will be used within the Dutch subsidiary.