《德电咨询:2024以数据为中心的时代战略建议研究报告(英文版)(51页).pdf》由会员分享,可在线阅读,更多相关《德电咨询:2024以数据为中心的时代战略建议研究报告(英文版)(51页).pdf(51页珍藏版)》请在三个皮匠报告上搜索。
1、Point of View Data-Centric CompanyDetecon Point of ViewData Centric CompanyTable of Contents1 Executive Summary 32 Introduction 52.1 What Are Data?52.2 Categories of Data 62.3 Three Levels of Data Integration in Companies 62.4 Challenges and Best Practices of Becoming Data-Centric 82.5 Launching the
2、 Process of Building a Data-centric Organization 92.6 Recommendations 93 Strategy in a Data-centric Age 103.1 Strategic Relevance of Data 103.2 Strategy in Data-centric Companies 113.3 What to Include in a Great Data Strategy 133.4 Data-centric Approach to Business Strategy D evelopment 153.5 Recomm
3、endations for Strategy in a Data-centric Age 164 How Data-centricity Improves Market Position 174.1 Data-centricity:An Increasingly Critical Factor in Market Competition 174.2 A Widening Gap:How Data-centric Market Leaders Gain a Decisive Competitive Edge 174.3 Recommendations for Marketing in the D
4、ata-centric Age 205 Data-centric Corporate Functions 245.1 Finance 245.2 Human Resources 265.3 Research and Development 285.4 Recommendations 296 Technical prerequisites for Data-Centricity 316.1 Autonomy and Data Democratization in a Data-centric Enterprise 316.2 The Infrastructure Prerequisites fo
5、r a Data-centric Company 376.3 Recommendations 397 Compliance 407.1 Introduction 407.2 Application of Regulatory Statutes Poses Major Challenges 407.3 Development of a Compliance Framework 427.4 Challenges 447.5 Result 447.6 Recommendations User-centric Approach to the Legally Compliant Use of Techn
6、ologies 44Page 2Detecon Point of ViewData Centric CompanyTable of Contents1 The technology section in chapter 3 will touch briefly on the means for securing the highest data value.2 Detecon Research3 https:/hbr.org/resources/pdfs/comm/workday/FinanceData.pdf4 How Much Data Is Created Every Day?27 Po
7、werful Stats|SeedScientificData serve as the foundation for all decision-making processes that,freed from the confines of silo views and with a 360-degree perspective,can improve prod-ucts and services and even result in completely data-based business models.Past research has demonstrat-ed that data
8、 centricity leads to better performance and productivity in tandem with higher market value.More-over,the quality and speed of the decision-making process,enriched by higher development productivity and customer insights,can attain a higher level.Todays companies find the wealth of the data availabl
9、e to them growing at an exponential rate as the collection of the information never ceases.1 Yet the number of DAX companies(seven of the 40)that have designated a position of chief data officer(“CDO”)for the manage-ment of their data remains low.2 Data strategy begins with a data-centric mindset an
10、d incorporates data and analytics into all corporate functions;the companys ability to meet constantly changing business and market 1 Executive Summary conditions in the data-centric age is sharpened and value for the business as a whole is generated.Becom-ing a truly data-centric company demands an
11、 emphasis on data as the core element of all operations;they must be regarded the starting point for making decisions and the optimization of processes and essential compo-nents of new products and services.As of this moment,comprehensive data-centricity can translate into a competitive advantage,al
12、though the scope can vary from one industry to the next.Still,it is now a critical factor in market competition and a key factor in future business models.Marketing activities such as price,promotion,place,and product are being adapted to assure data-based optimization of services in various industr
13、ies.Data-centricity will enable compa-nies to identify fields where they can obtain meaningful insights that will make a significant difference in their market placement and customer base.Finance departments pursuing a data-centric approach establish a position of strength as they are able to as-ses
14、s data value and share the insights and recommen-dations they have derived from the data with senior management and other departments within the organi-zation,staking out their role as custodians of enterprise data and the ultimate authority for analytics.3 The rise of big data and the advent of oth
15、er data-related technolo-gies in the past decade have brought about significant changes for human resources departments as well.The amount of information related to employee activities in an organization and stored by Google,Microsoft,Ama-zon,and Facebook has reached a staggering volume of 1,200 pet
16、abytes.4 R&D,with its focus on process inno-vation development,the updating of existing products,1,200 petabytes of information.1 Executive Summary Detecon Point of ViewData Centric CompanyPage 3and new product research and development,is another department enjoying huge potential.Redefining the R&D
17、 process in a dynamic decision loop that directly analyz-es the market/customer data for updates to the product/service is not possible without a data-centric approach in R&D.The democratization of data access inevitably demands changes in the way people think about the handling of data.In short,dat
18、a should be treated as a product that is ready for use and reliable.The data mesh platform is an internationally designed and disseminated data archi-tecture with centralized governance and standardiza-tion to ensure interoperability and a shared and harmo-nized self-serve data infrastructure.Rather
19、 than apply-ing top-down decisions regarding the formats in which data are stored for future users,the information is stored in its original form so that future users are com-pletely free to decide what transformations will best serve their needs.The solution desired by most organi-zations is a one-
20、stop platform for the performance of analytics and the development of meaningful insights without that do not require the assistance of a central IT team.The central IT team still exists,but it does not own the data.The general impression appears to be that strict laws would cause Europe to fall far
21、 behind China and the United States as such regulation would prevent the data that are or could be produced from being used for data-driven innovations.The assumption is that there would be few data-driven innovations coming out of Europe in future.Detecons analysis has identified three problem area
22、s for companies that have had to abandon data-driven innovation projects involving new technologies:1.The impact and requirements of the applicable data protection regulations were not included in the devel-opment from the outset.2.Internal compliance processes are found to be static,complicated,and
23、 time-consuming and require ex-tensive resources,sometimes from outside sources,for audits and consulting services.3.Some companies prohibit as a general principle the use of certain technologies such as machine learning or the exploitation of their own data because of the complexity and ambiguity o
24、f interpretations.Authorities have only recently begun to enforce fully the penalties established in the GDPR,actions that can have far-reaching consequences for both companies and individuals.Whatever approach is taken,we believe that the use of implementation and translation tools for laws and reg
25、ulations will enable users to develop legally compliant,data-driven innovations that will create future security for entrepreneurial activities.The solution most companies want is a one-stop platform.1 Executive Summary Detecon Point of ViewPage 4Data Centric Company5(Ackoff,1989,p.3)Ackoff,R.L.(198
26、9).From data to wisdom.Journal of applied systems analysis,16(1),39.6(Ackoff,1989,p.3;Kitchin,2014,p.910)Ackoff,R.L.(1989).From data to wisdom.Journal of applied systems analysis,16(1),39.AND Kitchin,R.(2014).The data revolution:Big data,open data,data infrastructures and their consequences.Los Ange
27、les:SAGE Publications.7 Adaption from Rowley(2007,p.164)Rowley,J.(2007).The wisdom hierarchy:representations of the DIKW hierarchy.Journal of information science,33(2),163180.2.1 What Are Data?The Oxford Learners Dictionaries(n.d.)define data as“facts or information,especially when examined and used
28、 to find out things or to make decisions.”The data-information-knowledge-wisdom hierarchy(DIKW)depicts data as the bottom level of the pyramid.The processing of the data produces information.5In turn,the analysis or interpretation of information cre-ates knowledge that,when applied,results in wisdom
29、.6 Naturally,all levels can and should be generated to one degree or another in every company.Companies have acquired enormous amounts of data that can prove to be useful and valuable,but they must learn how the conduct of analyses or modeling can identify the use and value generation from this stor
30、e-house.2 Introduction Figure 1:Levels of the DIKW Hierarchy7DataInformationKnowledgeWisdomDetecon Point of View2 IntroductionData Centric CompanyPage 58(Schloo,2021,p.43)Schloo,B.(2021).Data-centricity:A framework approach to assess the value of data for companies(Master Thesis,htw University of Ap
31、plied Science).9(Berndtsson et al.,2018,p.1)Berndtsson,M.,Forsberg,D.,Stein,D.,&Svahn,T.(2018).Becoming a data-driven organisation.In 26th European Conference on Information Systems(ECIS2018),Portsmouth,United Kingdom,June 2328,2018.10(Berndtsson et al.,2018,p.1).Berndtsson,M.,Forsberg,D.,Stein,D.,&
32、Svahn,T.(2018).Becoming a data-driven organisation.In 26th European Conference on Information Systems(ECIS2018),Portsmouth,United Kingdom,June 23 28,2018.2.2 Categories of DataData can be categorized from a company perspec-tive according to the cate-gories shown on the right.2.3 Three Levels of Data
33、 Integration in CompaniesSchloo8 identifies three levels of data integration i.e.,data-informed,data-driven,data-centric.The least complex level is data-informed as it uses solely descriptive analytics based on historical data9 for the creation of dash-boards.There is no central data strategy or the
34、 use of data science to carry out advanced analytics such as predictive,prescrip-tive analytics.10 Figure 2:Categories of DataBased on sourcesCaptured Captured data are directly collected when an activity is conducted in the real world or online,regardless of whether the purple of the collection is
35、to gather data about a particular topic.Examples include machine sensors,photo or video cameras(these raw data sources are fully unprocessed),credit scores the number of cars passing through a junction every hour are exam-ples.Derived Derived data are formed by performing some kind of pro-cessing ac
36、tion,on the data such as,aggregating or merging it.Quantitative Quantitative data is numerical and can be presented in a table.Qualitative Text,video,audio and photos are examples of qualitative data that can be turned in two Quan-titative data via classifications.Based on formsStructured Data that
37、is structured is con-sistent and may be saved and shown in a tabular style such as a data model or database and can be easily visualized and analyzed.Examples are names,addresses,etc.Semi-structured Semi-structured data has same structural patterns that may be seen when looking at its fields and sem
38、antics making it easy to order them.XML HTML and JSON files are examples of this kind of data.Unstructured Social media posts,digital notes&documents,photos,audio,and video files are common unstructured data.Based on structuresMeta Metadata is used to describe other data such as names,de-scriptions,
39、etc.Metadata helps users to understand how the other data is composed.Master Master data includes data about something which is essential to the companys operations such as employee,supplier,product,or customer data.Inventory Inventory data is changing con-tinuously and more often be-cause it descri
40、bes the current value and quantity of the com-panys inventories.Transactional Transactional data results from transactions of the business processes such as client orders or invoices and are hence the data type with the highest quan-tity and most changes.Reference The smallest data type is refer-enc
41、e data because it is generat-ed externally and serves the sole purpose to define elements universally consistent,like coun-try,currency,or language codes.Based on typesDetecon Point of ViewData Centric Company2 IntroductionPage 611 Schloo(2021,p.43)Schloo,B.(2021).Data-centricity:A framework approac
42、h to assess the value of data for companies(Master Thesis,htw University of Applied Science).A data-driven approach utilizes advanced analytics(predictive and prescriptive)and has established or is in the process of establishing a data strategy.The prin-cipal idea behind data-driven integration is t
43、he use of data for making decisions with the aid of an overarch-ing data strategy and data science.Data-centricity is the final and highest level of data integration,the stage at which data are the heart and soul of everything.In a data-centric system,data serve as the foundation for all decision-ma
44、king pro-cesses that,freed from the confines of silo views and with a 360-degree perspective,can improve products and services and even result in completely data-based business models.From a technical perspective,this implies the existence of an architecture that is independ-ent of the application a
45、nd that collates all the data into asingle model and the same format so that they can beused and combined easily.From an organizational perspective,the transformation to data-centricity is a change process during which employees must be moti-vated,educated,and trained to support the establish-ment o
46、f a data-centric culture and mindset.Figure 3:Level of Data Integration in the Company11Data-Informed Descriptive analytics Little data-based decision makingLevel 3Level 2Level 1Data-Driven Data science including advanced analytics Data strategy Application-centricity and data silos Siloed data-base
47、d mindset Mostly data-based decision makingData-Centric Data at the center of everything Necessary application-independence(no relevant data silos)Data-based culture and mindset Data access and usability for all employees Solely data-based decision makingPage 7Detecon Point of ViewData Centric Compa
48、ny2 Introduction2.4 Challenges and Best Practices of Becoming Data-CentricSpecific examples the challenges that may arise are available to support com-panies during the process of their trans-formation into data-centric operations.Experts speak of grouping these chal-lenges into two categories:“data
49、 and general technical challenges”and“organ-izational challenges.”The lessons learned and best practices mentioned by experts as success fac-tors during the process of transformation into a data-centric company can be grouped into general,technical,and organizational lessons learned.The challenges,l
50、essons learned,and best practices are summarized in the table on the right.CategoryChallengesLessons learned and best practicesGeneralShared data are not equivalent to equally accessible data identification of easy-to-use mechanisms applied by engineers to make data visible,accessible,and usable whi
51、le complying with data privacy regulations and avoid-ing bottlenecks Financial and personnel resources Transformation is a process,not a short-term project Establish an overall strategy/purpose and work,use case by use case;if an approach proves to be unsuccessful,drop it “fail fast and fail often”T
52、echnical Lack of knowledge about existing data Insufficient data and/or poor data quality Identification of valuable data objects Efficient processes for storage,processing,and sharingof masses of data without any loss of speed inreal-time data activities Diversity of data types and systems within a
53、 company incombination with varying requirements for data confidentiality Curated data teams and datasets to generate the same KPI values while avoiding duplication of activities and enabling self-service analytics Data catalog creating transparency about what exists where Save only relevant data as
54、 storage and maintenance cost money Invest in the harmonization of important data Data collection must be automated Automate compliant data access and deletion By default,data must be available to everyone.If not the case,explanation is requiredOrganizational Organizational/mindset silos and change
55、process Employees resentment or rejection of the change Employees fear loss of their jobs because of redun-dancy,especially when operations are automated Trusting the results of the data Power plays of individuals/departments unwilling to share the data Persuading all management levels to support th
56、e transformation Finding and training product owners and the links between the business and technical teams Centralized data governance and data ownership Need to involve all levels from the top in support of the overall strategy and from the bottom to secure full acceptance and obtain detailed insi
57、ghts about all units Identify pioneers(individuals from all units)so that their input and lessons learned can be utilized and their whole-hearted support is secured;they can be used to disseminate the full process through the entire organization Close and active collaboration among teams of business
58、 and technical employees Build-up of internal knowledge and simultaneously external(technical)expertise(ifnotalready available)to accelerate the process Generate the willingness to make the general mindset change and to share data by demonstrating the specific benefits for the individuals and/or how
59、 they will be able to perform their duties and responsibilities more effectively Promote the use of data and the related tools by making their use indispensable Table 1:Challenges and Lessons Learned and Best Practices of Becoming Data-centric1212 Schloo(2021,p.50)Schloo,B.(2021).Data-centricity:A f
60、ramework approach to assess the value of data for companies(Master Thesis,htw University of Applied Science).Page 8Detecon Point of ViewData Centric Company2 Introduction2.5 Launching the Process of Building a Data-centric OrganizationPast research has demonstrated that data centricity leads to bett
61、er performance and productivity in tandem with higher market value.Moreover,the quality and speed of the decision-making process,enriched by higher development productivity and customer insights,can attain a higher level.What is more,data-centricity lowers hardware and software requirements by elimi
62、-nating unnecessary applications,reduces the time and cost of integration data from new sources or a combi-nation of various sources,and puts an end to duplica-tion of the same work.Essentially,any business that wants to become data-cen-tric must consider various aspects that can be broadly categori
63、zed as business value or strategic,technical,and organizational concerns.Above all,it means gener-ating business value by placing data at the center,creat-ing a technical infrastructure that enables all employees to make data-based decisions from sharing and using the data,and promoting an organizat
64、ional environment in which the way employees think is guided by a data-centric mindset.Implementing data-centricity to some degree opens the door to the use and exploitation of company data.The factors that should be considered for the identifi-cation of potential value are economic,environmental,an
65、d social in nature while the ultimate goal concerns the economic benefits:increased profits and revenues,cost savings,and employee and customer satisfaction.In this respect,the companys vision and strategy should act as the underlying guiding principles and ensure that the identified use cases actua
66、lly support the goals the company is striving to achieve.2.6 Recommendations Shared data are not equivalent to accessible data;establish an overall strategy/purpose and work,use case by use case;if an approach proves to be unsuc-cessful,drop it “fail fast and fail often”.Create a data catalog that c
67、reates transparency regarding what exists where.Curate data teams and datasets to generate the same KPI values,avoid duplication of activities,and to pro-vide self-service analytics.Data collection must be automated.Automate compliant data access and deletion.Establish centralized data governance an
68、d decentralized data ownership.Identify pioneers(individuals from all units)so that their input and lessons learned can be utilized and their whole-hearted support is secured;they can be used to disseminate the full process through the entire organization.Build-up of internal knowledge and simultane
69、ously external(technical)expertise(if not already available)to accelerate the process.Generate the willingness to make the general mindset change and to share data by demonstrating the spe-cific benefits for the individuals and/or how they will be able to perform their duties and responsibilities mo
70、re effectively.Data collection must be automated.Page 9Detecon Point of ViewData Centric Company2 Introduction13(Gr&Spiekermann,2020)https:/www.isst.fraunhofer.de/content/dam/isst-neu/documents/Publikationen/Datenwirtschaft/ISST-Report/Fraunhofer_ISST-Report_Data-Strategy-Praxis-Report.pdf14(Forrest
71、er,2020)Data-centric businesses are 58%more likely to exceed revenue goals|TechRepublic3.1 Strategic Relevance of DataCan it be a coincidence that no fewer than four of the five most valuable companies are established and rec-ognized players in the digital world who exploit data as an important stra
72、tegic asset and a key value driver?Hardly!No,we at Detecon,along with investors around the globe and many researchers,13 firmly believe that data-centricity,especially the use of data in strategic decision-making and the creation of value through data,lays the foundation for long-term and sustainabl
73、e market success and a competitive advantage.A survey of more than 900 global business analysts found that 84 percent of the respondents believed it is essential to anchor key business decisions and strategies on the bedrock of data,yet half of them also indicated that their organizations do not at
74、this time routinely use data to drive decisions and strategies.14In Germany,just over half(22 out of 40)of the compa-nies listed on the DAX have appointed chief digital(rarely:chief information)officers with the designated 3 Strategy in a Data-centric Age Figure 4:The Evolution of Data(adapted from:
75、Gr&Spiekermann,2020)Data as products and servicesData as enabler of productsData as enabler of processesData as a result of processesData as enabler of strategyFor example,support in retrieving the current number in stockFor example,order-to-cash within business processesFor example,self-tracking de
76、vicesFor example,data marketplaces For example,anticipate market dynamicsAdded Value TimePage 10Detecon Point of ViewData Centric Company3 Strategy in a Data-centric Age15 Detecon Research,202216 SAP,2020role of helping to steer the organization through the digi-tal transformation.The presence of su
77、ch an officer is a good start,but the effectiveness of the position will be limited until data-centricity has been established throughout the entire company.Even fewer DAX com-panies(seven of the 40)have appointed a designated chief data officer(“CDO”)for the strategic management of data.15 SAP is o
78、ne example of a company that argues strongly in favor of investing in such a position and for the CDO to be a“strategic decision for managing data effectively for business growth,realizing efficiencies,and minimizing risk.”16Even though not all companies have reached the same level of maturity,techn
79、ological progress has increased the potential of data to create value over the course of time as is illustrated in Figure 4.The Detecon point of view:In the years ahead,the leaders of data-centricity will dominate their industries and build up a sustainable competitive advantage while companies negl
80、ecting data-centricity will be left behind.In view of this inevitable conclusion,developing an excellent data strategy and realizing it throughout your organization as the means of driving the generation of added value is essential if your company is to be steered to success!The transformation into
81、a data-centric enter-prise that also utilizes data as an enabler of the general business strategy is a key strategic decision demanding both top-level and company-wide commitment.Now is the time to make data a strategic issue!data strategy promotes the effective use of data in all business units and
82、 serves to drive the generation of added value,putting into companies hands the means to achieve data leadership and the related competitive edge.The second stage is the specific use of data as an enabler of the business strategy,a process also known as strategic analytics.A business strategy is dec
83、isive for Figure 5:Strategy in Data-centric Companies 3.2 Strategy in Data-centric CompaniesAt its very core,every strategy is about winning the game so in our view,companies that ne-glect data cannot claim to be pursuing a winning strategy.The strategy in data-centric com-panies evolves in two prin
84、cipal ways,both of which push the generation of added value to higher levels(see Figure 5).The first essential step is to draw up a clear data strategy that is based on the outstanding goals of the business strategy.The GoalsData-Centric Business StrategyStrategyDataEnabler/Strategic AnalyticsDataAl
85、ignment Functional Strategies(Sales,Marketing,Supply Chain,etc.)Divisional Strategies(Business Areas,etc.)IT StrategyPage 11Detecon Point of ViewData Centric Company3 Strategy in a Data-centric Age17(Atsmon,2016)https:/ Packard Enterprise,2021)https:/ the business case for a chief data officer|MIT S
86、loangrowth,positioning,and focus,making it the primary driver of company-wide generation of added value.A data-centric business strategy is guided by strategic considerations based on decisions powered by data.Leveraging data when making decisions encourages more informed directional guidance and re
87、duces the likelihood of bias causing errors in judgment,further facilitating generation of added value.As the effective reallocation of resources during the strategy process is in most cases the single most power driver of revenue growth,17 a data strategy and the allo-cation of resources relating t
88、o data management must be firmly embedded in and explicitly described in con-junction with the business strategy.Very few compa-nies,however,have reached the point of incorporating data into strategy.Below are the examples of two com-panies that have explicitly included data goals in their business
89、strategy:Allianz “Digital for us means enhanced data avail-ability and analytics.”18 Allianz recognizes data and analytics to be important drivers for future success with the overall aim of implementing a data-first decision-making culture.19 BASF “Digitalization is an integral part of our business.
90、We want to significantly improve the availability and quality of our process data.”20These goals stated in the business strategy are the start-ing point for the development of a data strategy that must subsequently be aligned with other strategies(functional,divisional,or IT)with the ultimate aim be
91、ing the establishment of a data-centric company.Yet organizations often fail to define a profound,distinc-tive,and far-reaching data strategy.21 The GAS region is still in an early stage of development when it comes to data strategies as indicated by a survey of executives in which 65 percent of the
92、 respondents noted that their companies do not have a data strategy,not even as one of the components of an IT strategy.22A study of more than 500 companies examining the use of data as an enabler of the business strategy found that organizations that treat data as a strategic asset and place them a
93、t the heart of the business strategy are more effec-tive when deducing relevant strategic insights,enhanc-ing business performance,and reshaping processes,products,and services.23 The logical conclusion is that top-level executives should pursue a strategic analytics approach to their business strat
94、egy,one in which data analytics informs key strategic decisions.The output of the data-centric business strategy can be monitored by the use of interactive dashboards providing real-time information about key metrics.Such features also im-prove business agility and the ability to initiate success-fu
95、l and proactive strategic moves.Detecon Point of ViewPage 12Data Centric Company3 Strategy in a Data-centric Age24(Ladley&Redman,2020)https:/hbr.org/2020/03/use-data-to-accelerate-your-business-strategy The Detecon point of view:A data strategy must define a timeline and the required resources in ad
96、dition to the technical and support conditions for its realization and be successfully integrated into the company.The data strategies currently found in many companies and the related activities are often too low-level,too short-term,and too far removed from overall business strategies to create a
97、data-centric enterprise.24 If optimal interplay be-tween business and data strategies is to become possi-ble,data can no longer be considered separately.More-over,in our VUCA world,it has become more important than ever before to make the right decisions quickly and readjust strategic directions whe
98、never necessary.This leads us to expect strategic analytics and value-gener-ating,data-centric business strategy to take their place as key success factors and competitive advantages for organizations in the coming years.3.3 What to Include in a Great Data StrategyA great data strategy begins with a
99、 data-centric mindset and integrates data and analytics into all corporate func-tions,laying the groundwork for the companys ability to meet constantly changing business and market condi-Figure 6:The Three Streams of Data Strategy Development Identify*gaps and issues:Resolving issues:data quality da
100、ta access/catalog data ownership/governance data enabling processes data security implementation roadmapStream 1:Create Business Value Stream 2:Empower Organization Stream 3:Enhance Data Engineering *(from streams 1 and 2 and additional analysis)Defne opportunities Defne focus areas Select initial u
101、se cases(23)Explore initial use cases Identify and assess data sources Create pilot(s)or PoCs Rollout&scale (multiple use cases)Identify potential Create joint vision and approach Identify capabilities and people Practical training(learning by doing)Build up skills&community Coaching sessions,know-h
102、ow transferPage 13Detecon Point of ViewData Centric Company3 Strategy in a Data-centric Agetions in the digital age and to generate added value throughout the enterprise.While not everyone in the company and beyond may automatically become enthusiastic about data-centric initiatives,the execu-tive l
103、eadership must instill a strong conviction of the necessity to launch the company on this transformative journey to data leadership.Success metrics for evalua-tion of the progress should be carefully considered with the aim of maintaining the motivation of all and of ensur-ing that the company is tr
104、uly benefiting from the data-centric initiatives.Another important point is to specify the required resources and the timeline for reviewing the data strategy.Starting from the goals derived from the data-enabled business strategy,the data strategy development pro-cess can focus on three key streams
105、 with the ultimate target of boosting value generation throughout the en-tire organization(see Figure 6).The first stream is the foundation for the business logic of the data strategy as its focus is on the creation of business value.Quite frequently,a key strategic issue when implanting a data stra
106、tegy is the lack of clarity about the business value and the resultant lack of sup-port from top management.The target in the“Create Business Value”stream is therefore to demonstrate clearly the business value of data.This is achieved by identifying specific opportunities and focus areas,ex-ploring
107、initial use cases,creating pilot projects of PoCs,and(in more advances phases)the rollout and scaling of multiple value-creating use cases,just to mention some examples.The second stream of the data strategy development seeks to empower the organization so that it can imple-ment the use of data acro
108、ss all sectors of its business and achieve the ultimate goal of enterprise-wide value creation.The“Empower Organization”stream is impor-tant for the design of a data strategy because it address key strategic issues,including the lack of data skills in the workforce,the lack of clarity regarding empl
109、oyee benefits of a data culture,and the general resistance to cultural change.In the long term,cultivating a data-cen-tric corporate culture presupposes the democratization of data intelligence by providing all employees not just developers or data scientists with the skills required to place data a
110、t the heart of decision-making and busi-ness operations.This organizational enablement can be achieved by means of strategic change management initiatives arising from a company-wide vision and ap-proach,initial capability assessments,and continuous skill development from the introduction of practic
111、al training programs,coaching sessions,and knowledge communities.The third stream “Enhance Data Engineering”seeks to improve technical capabilities and resolve key strate-gic issues such as inadequate data quality,a lack of data access,and the unavailability of processing and analyt-ics tool.The tec
112、hnology needed to attain the highest data value must be taken into account and the implementa-tion road map for its realization must be prepared.The full length of the data value chain must be reviewed:the collection,management,transformation,visualization,and storage of data that will support data-
113、centric value creation for the company.Data thinking is a useful and innovative framework that can be utilized during the development of a data strategy.The framework combines data science with design think-ing,so the focus is not solely on data-analytics technolo-Page 14Detecon Point of ViewData Ce
114、ntric Company3 Strategy in a Data-centric Age25(Ladley&Redman,2020)https:/hbr.org/2020/03/use-data-to-accelerate-your-business-strategy26(Mulligan et al.,2021)The strategy-analytics revolution|McKinsey27(Tableau,2022)https:/ Decision Making:A Primer for Beginners(northeastern.edu)29(Mulligan et al.,
115、2021)The strategy-analytics revolution|McKinsey30(Stobierski,2019)The Advantages of Data-Driven Decision-Making|HBS Onlinegies;it goes on to encompass the design of user-centric solutions with strategic business potential.Data thinking applies design thinking methods for the identification of custom
116、er needs and creative ways to find new solutions for the mastery of data-driven challenges.3.4 Data-centric Approach to Business Strategy D evelopmentBy and large,strategists have rarely made any use of data and advanced analytics for the development of the busi-ness strategy,25 and yet the right an
117、alyses offer potential for significant improvement of business strategies and strengthening of the executive leaderships position!26 A data-centric business strategy based on strategic analyt-ics offers tremendous opportunities for value creation and business agility to organizations.One major benef
118、it is the improvement of decision-mak-ing that is achieved by the use of a more objective and fact-based set of information.Data-centric decision-making utilizes data in support of strategic business deci-sions that are in alignment with the organizations goals,objectives,and initiatives.27 Two outs
119、tanding success stories from the digital business world could not have been realized without a data-centric decision-making approach and the data-centric business strategies de-veloped in conjunction with it.If they had not moved in this direction,Amazon would still be a simple online bookstore and
120、Netflix would still be mailing out DVDs that cannot even be played in many households today.28 Both of these companies,however,made key strategic business model decisions based on data and realized their organizational success by pursuing data-centric business strategies.Fast forward to today:many p
121、eople cannot even picture their everyday lives without these two companies.Strategic analytics can generally prove to be a useful tool for identifying trends that are still in their early stages or for revealing a variety of opportunities for business growth and strategic value creation.Intelligent
122、analyses includ-ing sentiment and network analysis of data from publicly accessible sources such as myriads of websites,com-pany descriptions,patent filings,news sources,clinical trial reports,geospatial,demographic,and M&A data,or academic papers,just to mention a few can uncov-er patterns pointing
123、 to emerging trends,go-to-market demands,or recognition in good time of changes in cus-tomer sentiment.29 What is more,these analyses can reveal new and attractive customer segments and ac-quisition targets,new applications for existing offerings,or even ideas for products,services,and new business
124、models.Starbucks,for instance,makes its decisions about the acquisition of locations in consideration of analyses of demographic and traffic data as a tool for ensuring high success rates for its investments.30 What is more,mathematical modeling and simulation can help to anticipate complex market d
125、ynamics by approx-imating real-world behavior.Agent-based modeling and Amazon would still be a simple online bookstore and Netflix would still be mailing out DVDs.Page 15Detecon Point of ViewData Centric Company3 Strategy in a Data-centric Age31(Mulligan et al.,2021)The strategy-analytics revolution
126、|McKinseysimulations,Monte Carlo analyses,and a variety of ma-chine learning approaches can be employed to realize differing scenarios of future situations and to assess their risks or to highlight important trade-offs and assumptions associated with a series of strategic choices.Methods of this typ
127、e can be especially useful whenever manage-rial intuition is inadequate to account fully for the impli-cations of the actions of multiple independent actors such as competitors or customers.31Additionally,the strategic analytics approach to business strategy frees companies to become more flexible r
128、esil-ient,proactive,and agile.By anticipating the strategic moves of their competition and becoming a first mover thanks to the timely identification of business oppor-tunities,companies pursuing a data-centric business strategy can lay a solid foundation for success and strengthen their competitive
129、 position on the market.Leveraging the capabilities for utilizing data in conjunc-tion with advanced analytics in strategy development for the radical improvement of strategic decision-making and of the overall business strategy will help any enter-prise to stay ahead of the game in this data-centri
130、c age.3.5 Recommendations for Strategy in a Data-centric AgeBecoming a data-centric company means placing data at the center of everything as they are the enablers of strategy development,the basis of decision-making and process optimization,and a part of new products and services.Indeed,the data in
131、ventory must become a strategic asset and key value driver for the organization.We have the following recommendations for companies wanting to establish this prerequisite for long-term and sustainable market success and competitive advantage.(1)Explicitly include data as a value driver when deliber-
132、ating possible strategies;secure a clear demonstration,externally and internally,from the top level management that the organization is on a journey to becoming a data-centric company.(2)Develop a convincing data strategy(taking into ac-count the three streams“Create Business Value”,“Em-power Organi
133、zation”,and“Enhance Data Engineering”)that promotes the effective use of data in all business units and positions to drive company-wide value creation.(3)Employ strategic analytics in data-centric business strategy development while leveraging data-powered decision-making to obtain more informed dir
134、ectional guidance,improving value creation and business agility.(4)Keep in mind that a data strategy must define a time-line,the required resources,and the technical and sup-port demands for its realization and successful integra-tion into the companys operations.The data strategies and their activi
135、ties currently found in many organizations are often too low-level,too short-term,and too discon-nected from the general business strategies to create a data-centric enterprise.Data can no longer be consid-ered in isolation if optimal interplay between business and data strategies is to be achieved.
136、(5)Data thinking is a useful and innovative framework that can be effectively utilized during the development of a data strategy.The framework combines data science with data thinking with the consequence that the focus is not solely on data analytics technologies;instead,it extends to the design of
137、 user-centric solutions with strategic business potential.Page 16Detecon Point of ViewData Centric Company3 Strategy in a Data-centric Age4.1 Data-centricity:An Increasingly Critical Factor in Market Competition While,as briefly described above,the inclusion of data as a key starting point during st
138、rategic development is becoming increasingly inescapable,they have already achieved a position of crucial relevance and practical at the interface between companies,their customers,and their competitors in day-to-day market competition.Companies are desperately seeking to surpass their competitors b
139、y enhancing their skills in analyzing cus-tomer data and exploiting them for marketing purposes.By taking a data-centric approach,they can significantly expand the range and heighten the quality of their in-sights for the optimization of their own products and to expand their business model through
140、the introduction of additional services and products complementing their current portfolio onto the market.This chapter will explore the means by which marketing can capitalize even further on data-centricity and discuss the key factors that must be considered when moving in the direction of a data-
141、centric business model.4 How Data-centricity Improves Market Position4.2 A Widening Gap:How Data-centric Market Leaders Gain a Decisive Competitive EdgeIn many industries,real-time analytics and optimization have long since become standard practice rather than a novelty.Data-centric optimizations ar
142、e applied exten-sively across all aspects of marketing.Figure 7:Data-centric Approach to 4P Data-driven value-adding services Advanced site-specifc optimizations Cross channel improvementsData-informed product development Data based product optimization Seamless product&service feedbackCustomized co
143、ntents and ofers Micro-targeting Campaign efciencyReal-time pricing Customized discountsDynamic pricing strategyProductPricePlacePromotionData-centric 4PPage 17Detecon Point of ViewData Centric Company4 How Data-centricity Improves Market Position PricePricing is a simple example of the use of data-
144、based optimization that is especially prevalent in the B2C sec-tor.Real-time dynamic pricing based on data for factors such as competition,consumption trends,and similar sources is widespread.Amazon alone makes several million price adjustments on its own platform every day.Surprising perhaps,techno
145、logy companies were not the first adopters of dynamic pricing;that honor goes to the hospitality and travel industry(especially airlines as their core mission is to market each flight at the highest possible capacity while maximizing the profit margin per seat).In the meantime,virtually all large re
146、tail companies are moving to adopt dynamic pricing as a means of maintaining their competitiveness and optimizing mar-gins.Trading and comparison platforms have already increased significantly the level of transparency on many markets.Todays customers are able to compare prices in only seconds,so su
147、ppliers are under even more pressure to set prices appropriately and adapt them to changing market conditions.Even the offline retail sector is adopting dynamic pricing optimization that has become possible with the spread of digital price tags.Suppliers will certainly continue to make use of dynami
148、c pricing schemes so as to keep up with the pace of activ-ities on their particular markets.Nevertheless,they should be aware of the challenges of dynamic pricing models that have become apparent in recent years,especially if they want to prevent significant harm to repu-tation and loss of trust in
149、their brands.The ride services Uber and Lyft are warning examples;they were massively criticized publicly when,following an assassination at-tempt,the demand for rides in the impacted neighbor-hood skyrocketed and prices were dynamically raised to five times their previous level.Reaping the rewards
150、of data-based pricing requires companies to give special consideration to the unique characteristics of the mar-ket on which they operate and to brand perception.Ulti-mately,they must be prepared to refine their data-based pricing policies and to be more sensitive and consider-ate in the future.Prom
151、otionCustomer-related data analysis is essential to a modern company to much the same degree as competition-ori-ented data analysis.Companies draw on a broad range of disparate data sources so as to obtain an assessment of potential customers and their needs that is as accu-rate as possible and so a
152、s to tailor more specifically both their product line and their marketing approach.Person-alized content is more effective at reaching customers and can significantly raise the conversion rate as well.More in-depth knowledge of customers translates into a more efficient use of marketing budgets.An i
153、mpressive example of the opportunities offered by data centricity is the Deutsche Bahns“No Need to Fly”marketing campaign.The starting point for the campaign was the determination of the international travel destina-tions attracting the interest of individual Facebook users.Photos of German destinat
154、ions similar to the preferred international sites were selected and the air and rail fares to both destinations form the specific users location were calculated.The comparison of the two destinations were subsequently displayed to the user as an advertise-ment.This data-centric approach posted a cli
155、ck-through Todays customers are able to compare prices in only seconds.Page 18Detecon Point of ViewData Centric Company4 How Data-centricity Improves Market Positionrate that nearly quadruple that of other campaigns,and the cost per click of the campaign was reduced by 59 percent in comparison with
156、previous campaigns.This example illustrates the potential to be found in making greater use of data-centric approaches in marketing for achieving both greater cost efficiency and heightened impact.The relevance of customer data in marketing,already at a high level,became obvious when iPhone manufact
157、urer Apple introduced a new opt-in model for the analysis of user data by app and website providers in spring 2021.Since the introduction of this feature known as“App Tracking Transparency,”every app must obtain the users consent to the analysis of his or her data for marketing purposes.Since a larg
158、e proportion of users did not con-sent to an analysis of their personal customer data,the targeting options of many advertising campaigns were significantly restricted and the effectiveness of the cam-paigns declined.The result of the introduction of the new feature was an almost immediate loss of r
159、evenues totaling$10 billion for the major platform operators Facebook,Twitter,Schnapchat,and YouTube in 2021.PlaceAside from the optimization of advertisements and“customers also bought”placements,some data-based services present clear added value for customers and opportunities for differentiation
160、from competitors.One example of this aspect is sizing for online clothing retail-ers.More and more retailers have begun to customize their automated recommendations for individual cus-tomers on the basis of the measurements previously provided by the latter.These recommendation models are further re
161、fined by the inclusion of extensive data-sets comprising previous orders and returns from all customers to identify more precisely what sizes will best fit specific customers.Armed with this information,cus-tomers can make more informed purchasing decisions.The greater confidence in product selectio
162、n both im-proves the conversion rate and reduces the high vol-ume of returns prevalent in online fashion retailing.Differentiation through data-based services of this type is of high strategic relevance for e-commerce.The chal-lenge for providers is found in the acquisition of the large volumes of d
163、ata required to refine the recommendation models.Large retail brands and platform operators like Zalando or Otto have the reach essential for the devel-opment of these types of services that draws on their own customer data;they are in a position to develop pro-prietary solutions for their own platf
164、orms.Smaller stores and fashion brands rely instead on the integration of platform-independent technology providers such as Presize.32 These platform providers give smaller retailers the chance to compete with larger platforms because they maintain powerful recommendation models trained with the dat
165、a from all participating online stores.This example from the fashion retail trade demonstrates how companies can significantly improve their customer experience by analyzing the customer data they already possess.This opportunity is not restricted to retailers with high transaction volumes.Depending
166、 on the use case,companies may find it in their interest to share data with partners and competitors for the development of new services for their own customers and the strengthening of their own data ecosystem.32 https:/textile-network.de/de/Fashion/CAD-CAM/Mega-deal-Presize-start-up-owned-by-Meta-
167、nowPage 19Detecon Point of ViewData Centric Company4 How Data-centricity Improves Market Position ProductProduct development can also be optimized significantly through data centricity based on collected user data.Netflix is a good example.Thanks to its enormous user base,the streaming service can a
168、nalyze its users pref-erences with extreme precision.This analysis of user behavior is not limited simply to the programs that are watched;it reveals what scenes were repeated or skipped or the point at which users interrupted the streaming.Netflix uses the obtained data to integrate the more attrac
169、tive elements into its new films and series,match-ing customers interests with a high degree of probability.The evaluation of user behavior at this level of detail represents a significant competitive edge over tradi-tional television broadcasters.This is a prime example illustrating that data use i
170、s not limited to only one link in the value chain.On the contrary,the data from one stage of the value chain can generate extensive poten-tial for added value at other stages.The lesson is that companies should attempt to collect data along the full length of the value chain,including,if possible,th
171、e up-stream and downstream links from suppliers to cus-tomers.The examples above are representative of the broad range of opportunities arising from data centricity at all levels of marketing,from the optimization of prices to personalized customer targeting,data-based add-on services,and product de
172、velopment and optimization.They encourage and remind companies to make the most of the data available to them.As of the moment,com-prehensive data centricity can result in a competitive advantage;as its adaptation expands across a growing Companies should attempt to collect data along the full lengt
173、h of the value chain.number of markets,data optimization will become a critical element for survival in a competitive environ-ment.4.3 Recommendations for Marketing in the Data-centric AgeAlthough the applications of data-driven insights into marketing activities are impressive,this facet of data ut
174、ili-zation will ultimately be inadequate for many traditional business models.Indeed,traditional corporations will be forced to take even more dedicated and courageous steps to data centricity going well beyond the superficial flagship initiatives still being launched by many corpora-tions(and that
175、have their roots in a pre-digital era)be-cause these efforts often stall after the development of a digitalized minimal viable product(MVP).They inevita-bly benefit no more than a small fraction of the compa-nys current products and value creations.This next section proposes three major steps that w
176、ill reap the full benefits from the use of data in sustaining,stabilizing,and growing business on highly competitive markets.Page 20Detecon Point of ViewData Centric Company4 How Data-centricity Improves Market Position1.Start your transition to a data-centric business model nowThe initial actions f
177、or digitalization of a business and its current portfolio are(in the best case)based on the primary business data obtained from previous opera-tions.Data-informed activities of this nature might aid in handling the most crucial short-term challenges encoun-tered during stabilization of the business.
178、They are only the first steps on the journey to data centricity,however;this move to data-informed decision-making based on previous business data will not secure a front-runner position for companies still operating with traditional business models whether for product advancement or customer knowle
179、dge or on the market.This shortcom-ing will become especially evident when a comparison is made with new digital competitors who are building robust platforms,forging strong partnerships,and extend-ing their value chains.Successfully competing with new digital players demands of the corporate leader
180、s that they leave the organizations comfort zone,broaden their view of the supply chain in its entirety,and explore new methods for engaging with customers and gaining deeper insights from available data.By taking these steps,companies will be able to identify the fields from which meaningful conclu
181、sions that will make a significant difference in their market placement and clientele can be drawn.2.Stay in motion on your path to a fully data-centric business modelMerely adapting the methods used by traditional compa-nies to create value,however,is too narrowly focused as an approach to achieve
182、the goal of becoming a fully data-centric company on the market.Data-driven platform business models(Amazon,Airbnb,Alphabet on the international stage;Check24 in the GAS region)create value from the data gathered and analyzed within their platforms,but not rely solely on the data acquired during the
183、ir own interaction with customers.They also integrate the data from the partners into the develop-ment of insights that enable them to tailor offerings and establish additional revenue streams.Although this oft-quoted data-centric business model is not by any means the ideal business model approach
184、for all companies,it should be kept in mind as an image of the enormous potential of data collected by businesses and their part-ners along the value chain.3.Find an approach that is yours only,based on your unique identity and dataFortunately,the data-driven platform model is not the final word;the
185、re are a minimum of two other types of data-centric business models of proven effectiveness in safeguarding or even growing revenue streams sub-stantially across most traditional industries.Data enhanced products and services building on existing products and services Data monetarization obtaining s
186、ubstantial and quan-tifiable benefit from own and third-party dataFor instance,T-Mobile US has successfully completed a large-scale project for the integration of data from its multiple local sources into one central source involving the installation of an IT architecture compatible with data integr
187、ation,providing to T-Mobile the instruments for the optimization of their marketing activities addressing their selected target groups.T-Mobile has been able to generate additional revenues from tailored cross-and up-selling offerings that in turn aided in safeguarding existing revenues through the
188、implementation of a dedi-Page 21Detecon Point of ViewData Centric Company4 How Data-centricity Improves Market Positioncated churn management system used to identify key patters of customer dissatisfaction such as bad cover-age or contract terms unfavorable to customers.Thanks to all these actions,T
189、-Mobile has been able to cut its churn rate by more than half within the last decade.In contrast,American Express analyzed huge amounts of its own data and applied machine learning technol-ogy so that it could predict customer churn and initiate effective countermeasures.The company did not stop wit
190、h this type of data-centric value creation,however.Exploiting its position as a service provider to millions of companies,it launched a dedicated program called Amex Advance that could provide to its clients deeper insights concerning the behavior of the latters custom-ers.American Express itself be
191、nefits by reducing its Figure 8:Three Generic Types of Data-centric Business Models Data-enhanced products&servicesDatamonetarizationData-drivenplatformInformationInformationInformationProductrelatedCustomerrelatedEnterpriseCustomerProductRevenueSupportExperiences InformationRevenue ProductRevenueCu
192、stomerRevenue(and/or information)InformationInfo-centricenterpriseInformationRevenueInfo-centricenterprise3 rd Party3 rd PartyPage 22Detecon Point of ViewData Centric Company4 How Data-centricity Improves Market Position Figure 9:Churn Rate of Postpaid Contracts at T-Mobile US from Sta
193、tista,2022own exposure to the constantly growing market of digital payment providers,pre-serving its above-market margins and premium price level while simultaneously gaining new customers for the additional data-driven services.In summation,the steps and the brief examples described above indicate
194、that a data-driven mindset is the best basis for establishing a durable competitive advantage.While larger amounts of data lead to deeper insights,any move to-wards data centricity should build on the strengths of your present products and services on the market as this will enhance the chances of l
195、ong-term success.In addition,the key role of the specific organizational capability to adapt and learn should never be underestimated as this is what determines the pace and the accompanying measures of the holistic transformation of organization.This will be described in greater depth in the fol-lo
196、wing chapter.0,0%Q1 10Q3 10Q1 11Q3 11Q1 12Q3 12Q1 13Q3 13Q1 14Q3 14Q1 15Q3 15Q1 16Q3 16Q1 17Q3 17Q1 18Q3 18Q1 19Q3 19Q1 20Q3 20Q1 21Q3 21Q1 220,5%1,0%1,5%2,0%2,5%3,0%3,5%Churn RatePage 23Detecon Point of ViewData Centric Company4 How Data-centricity Improves Market Position34 https:/www.fm- https:/h
197、br.org/resources/pdfs/comm/workday/FinanceData.pdf 36 https:/hbr.org/resources/pdfs/comm/workday/FinanceData.pdfCorporate functions and R&D are fundamental to most organizations,and Detecon views data centricity as an approach that can achieve a unique placement of corpo-rate functions and R&D,going
198、 beyond the mere compi-lation of data to bring together the decision-makers in the business and to provide to them the tools required to identify,optimize,and drive value.5.1 Finance The rising demand for data-related services concentrates above all on the supporting role of finance departments,who
199、are facing steadily more exacting requests for data related to pricing and trade and for risk and compliance analyses.34 Many finance teams,however(especially in large,sluggish conglomerates),are falling short in their attempts to satisfy these requests because of the low level of maturity of their
200、digital analytics.355 Data-centric Corporate Functions Detecon regards the following fundamental principles to be the most important pillars for companies looking to shift to a data-centric approach:Finance must be in the drivers seat for data value assessment.The use of advanced analytics in financ
201、e must be welcomed and insights that actually add value must be derived.The benefits of automation must be reaped and the finance departments position must refocus on value creation.5.1.1 Finance in the Drivers Seat for Data Value AssessmentAs data become an ever more significant factor for inno-vat
202、ion,competition,and the improvement of profitability in all industries,they have turned into a major asset,representing a value that organizations constantly seek to maximize.Finance executives are best qualified to answer questions about the value of data.A data-centric approach allocates this resp
203、onsibility to finance,positioning it to assess data value and share data-derived insights and recommendations with senior management and other departments within the organi-zation and to stake out its role as the custodians of enterprise data and the go-to authority for analytics.36 Ultimately,the b
204、est way to garner support from senior management is to demonstrate conclusively the value of data.Finance executives are best qualified to answer questions about the value of data.Detecon Point of ViewData Centric Company5 Data-centric Corporate FunctionsPage 2437 https:/ https:/ easyJet used big da
205、ta to achieve many millions of revenue improvements.StrataConf”Marketing Bluemetrix on June 1,2018.https:/ https:/hbr.org/resources/pdfs/comm/workday/FinanceData.pdfThe results of a recent study reveal that intangible assets(R&D,reputation,data,and the like)comprise approxi-mately 90 percent of S&P
206、500 market value.37 This figure alone raises the question of why organizations do not give adequate weight to these assets in the valuations.Due diligence in the handling of data should help to make more informed decisions.Obviously,the integration of due diligence when pro-cessing data in any comme
207、rcial or financial due diligence procedure is sorely lacking despite the crucial part it has to play,regardless of the investment or financing project that is under discussion.5.1.2 Embrace Advanced Analytics in Finance and Create Insights That Add ValueStaying up to speed with the current pace of t
208、echnologi-cal change is a daunting task,and there is certainly a temptation to cling to familiar working methods.As the masses of available data continue to grow,however,business leaders are seeking in-depth insights that will tie business activities to long-term value,facilitate the modeling of sce
209、narios in real time,and indicate how resources can be allocated efficiently.There must be a shift in focus from the provision of re-ports to the delivery of added-value finance services self-service analytics,spend analytics,cash flow fore-casting,and customer or product profitability projec-tions t
210、hat can offer proactive business guidance.38One such example is the British multinational airline group easyJet,which uses big data and advanced analytics to support its finance department in steering the pricing lifecycle,which involves the management individually of 500,000 flights and their ancil
211、lary services and 30,000 daily adjustments.easyJet has developed an algorithmic ecosystem incorporating historical flight performance,live analysis that can forecast and track demand pro-gress,and events analysis to handle the variations in demand during the year.395.1.3 Reap the Benefits of Automat
212、ion and Refocus the Finance Function on Value CreationAdvanced analytics opens the door to the benefits of automation.40 As machine learning and AI take over more and more of their routine tasks,finance profes-sionals gain additional time to refocus their tasks.In this sense,Detecon expects an incre
213、asing shift in the activi-ties of future data-centric finance functions from reac-tive responses,transactional activities,and retroactive analysis to more sophisticated,collaborative,and for-ward-looking procedures driven by the desire to gain insights.One example of this change can be seen in Xero,
214、a pro-vider of cloud-based finance and accounting applica-tions for small businesses,which offers accounting software based on computer capability for the simple posting of receipts and invoices.Moreover,it has intro-duced a new AI-based feature for bank account recon-ciliation that learns from mill
215、ions of historical transac-Detecon Point of ViewData Centric Company5 Data-centric Corporate FunctionsPage 2541“Citi Global Trade Uses AI to Digitize Compliance in Next Generational Project”.Citigroup on April 29,2019.https:/ How Much Data Is Created Every Day?27 Powerful Stats|SeedScientific43 Meet
216、 Robot Vera,the Latest AI Development to Hit Human Resources“.J.B.F.News.By Maria on May 2,2018.https:/ How Chatbots Help HR Managers Improve Employee Experience&Reduce Costs(botcore.ai)45 Chatbots expected to cut business costs by$8 billion by 2022()tions and further reduces manual data errors for
217、its busi-ness customers.Citi,a multinational investment bank and financial services corporation headquartered in New York City is another example;it uses AI to stream-line the time-consuming,highly manual processes of reviewing huge volumes of global trade transactions while ensuring regulatory comp
218、liance.415.2 Human Resources In the past,data have generally been regarded as irrele-vant for HR departments.Even if this were not the case,data did not play a strategic role,being used rather to calculate KPIs measuring person-day counts or absen-teeism while insights critical for the business with
219、in complex data were ignored.While many companies still use head-day counts to manage their employees sick leave,Previa,by contrast,has introduced a graph data-base that delves into the figures to uncover the hidden reasons causing absenteeism.Based on the results,companies can improve working condi
220、tions and foster the well-being of their workforce.The rise of big data and the advent of other data-related technologies have led to significant changes in the posi-tion of human resources in the last decade.Google,Microsoft,Amazon,and Facebook maintain storehous-es containing a mind-boggling 1,200
221、 petabytes of infor-mation42 that is exclusively related to employee activities in the organizations.Since HR is a people-oriented corporate function,we will examine how a data-driven approach could sup-port the core activities of the department related to people their recruitment,management,retenti
222、on,and development.5.2.1 Recruitment:Using Data to Find the Best ApplicantsWithout data,HR can do no more than make assump-tions about the background education,experience,or personality when assessing prospective employees,inserting an element of uncertainty into recruiting deci-sions.For instance,S
223、tafory,an award-winning start-up,has developed Robot Vera,an application that aids major companies in their recruiting of personnel.Available in various languages,it searches for suitable candidates,sends job descriptions to them,and conducts inter-views,reducing overall recruitment costs for the co
224、m-panies by 50 percent while working around the clock.435.2.2 Management:Using Data to Improve Organizational Efficiency and EffectivenessTotal operational costs,whether for recruitment or ad-ministrative tasks,can be substantially reduced by the automation of processes using chatbots.44 CNBC calcu-
225、lates that chatbots will reduce costs by as much as$0.70 per interaction and to cut business costs by a total of$8 billion as of 2022.45Telekom Romania has launched the chatbot ANA that automates the flows of internal HR activities and pro-vides self-service support for employees.It aids in the proc
226、essing of requests for leave and medical certifi-Detecon Point of ViewData Centric Company5 Data-centric Corporate FunctionsPage 2646“Telekom deploys DRUID AI chatbots to automate HR processes for 4000 employees.”Druid Enterprise Chatbots on May 28,2021.https:/ Intelligence Is for Real”.Tata Group i
227、n September 2019.https:/ North Star BlueScope Steel Taps IBM Watson and Wearable Devices to Monitor Activity of Workers in Extreme Environments()49 How Google Is Using People Analytics to Completely Reinvent HR TLNTcates,displays remaining days of leave,and updates em-ployee data using optical chara
228、cter recognition(OCR)technology.ANA also answers general HR questions,all the while perfecting and expanding the range of its per-formance by learning from direct interactions with its users.46 Yet another example is Tata Steel,a multina-tional steel company from India with a workforce of more than
229、30,000.It has developed the HR bot Cara,which answers routine questions on HR policy,saving valuable productive time and effort for the HR team.47 5.2.3 Retention:Using Data to Raise Employee Satisfaction and to Encourage Them to Stay with the Company LongerMeasurements of factors related to working
230、 conditions such as temperature,humidity,movements,and pulse rates can be used to determine whether the working conditions for personnel are sustainable.Seeking to guarantee the well-being of its workforce,North Star BlueScope Steel utilized the IBM Employee Wellness and Safety Solution for the prot
231、ection of their employees in extreme working environments.48 This appli-cation is a research project incorporating IBM Watsons cognitive computing power and sensors for the analysis of data collected from workers wearables and for the provision of data to North Star management in real time whenever
232、the technology detects potentially problem-atic conditions.Googles PiLab is a pioneer in innovative people manage-ment.It conducts practical experiments within Google to determine the most effective approaches for people management and the maintenance of a productive envi-ronment by raising employee
233、 satisfaction (including the type of reward that makes employees the happiest).The lab has even improved employee health by reducing the calorie intake of the personnel at eating facilities on the basis of scientific data and experiments(simply by reduc-ing the size of the plates).495.2.4 Developmen
234、t:Help Employees to Grow Together with the Company Using Data to Close the Gap Between the Current Competence Base and the Future Workforce MapHR departments carry out strategic workforce planning by aiding employees in the design of a cross-functional career path,in the discovery of opportunities f
235、or per-sonal promotion,or in the preparation for digital work situations,shifting from an operational to a strategic role for the long-term development of a company.The appeal of strategic workforce planning lies in its promise to adopt a data-driven approach to the structuring of the work-force pla
236、n from the Markov model to dynamic program-ming,from stochastic modeling to(meta)heuristics.NASA,for example,uses a graph database to map all information about employee skills,capacities,projects,and so on for the dynamic organization of the work-force and in preparation for future missions(e.g.,Mar
237、s or moon projects).In creating the body of information concerning the prerequisite skills for various occupa-tions,the associations team used a database called IBM Employee Wellness and Safety Solution for the protection of the employees.Page 27Detecon Point of ViewData Centric Company5 Data-centri
238、c Corporate Functions50 NASA reaches for graph DB to find people,skills for Moon and Mars missions The Register51 GSMA(2019).AI in Network.Use Cases in China.https:/ Who We Are|Enthought,Inc.53 Enthought Tool Suite Enthought Tool Suite documentationO*NET from the United States Department of Labor in
239、 conjunction with ESCO,the European Skills,Compe-tences,Qualifications,and Occupations database.505.3 Research and DevelopmentThe research and development(R&D)department fo-cuses on process innovation development(e.g.,thermal management solutions in the manufacturing sector),the updating of current
240、products,and research into and development of new products.Growing reliance on software and the availability of simu-lation and automation technologies have accelerated the innovation cycle.Corporations with a well-rounded R&D strategy are rapidly scaling up innovations that often threaten to upset
241、established business models or steer industry growth into new areas.Nonetheless,a tradi-tional,theory-centric R&D approach adhering to the traditional research process cannot maintain pace with the rapidity of changing market requirements.Only a data-centric approach in R&D can redefine the R&D proc
242、ess as a dynamic decision loop that directly ana-lyzes market/customer data for updates of the prod-uct/service.5.3.1 Ecosystem RebuildingBy pursuing a data-centric approach,a company or organization can proactively create an innovation eco-system,effectively leverage external resources globally,and
243、 readily translate research into innovative,market-ready products and services.Any new telecommunications network should be built in fulfillment of the current technical requirements of 5G.China Mobile,for instance,which boasted close to a billion mobile subscribers in 2020,has developed an AI R&D p
244、latform that provides comprehensive,high-qual-ity,and tagged AI training databases that can be shared throughout the entire company and that supports cen-tralized management of enormous quantities of multido-main data.China Mobiles goal is to build an AI ecosystem that will serve the network,market,
245、service security,management,and other areas.515.3.2 Enhance Agility and Speed in Product Innovation with Reduced Time-to-MarketA data-centric R&D department will actively aid in the management of a project portfolio to maximize ROI and systematically manage KPIs to steer the organization.Enthought,5
246、2 a digital solution provider focusing on sci-ence industries such as material science and chemistry,life science,semiconductors,and energy,has designed a suite of data analytical tools aimed at promoting digi-tally powered innovation for industries operating in sci-entific fields.The Enthought appr
247、oach merges aspects of skill development,software,technology,and digital consulting that accelerates consumer product revisions involving less trial and error through the use of machine learning.53 Only a data-centric approach in R&D can redefine the R&D process.Detecon Point of ViewData Centric Com
248、pany5 Data-centric Corporate FunctionsPage 2854 Jrjens,J.(2016)The Industrial Data Space:Digital Industrial Platform Across Value Chains in All Sectors Of The Economy.Web:https:/ec.europa.eu/futurium/en/system/files/ged/industrial_data_space.pdf55“Merck KGaA,Palantir Form Joint Venture to Develop Ca
249、ncer Data Platform”.Inside Precision Medicine on November 20,2018.https:/ symptoms search trends to inform COVID-19 research”Google Health.By Evgeniy Gabrilovich on September 2,2020.https:/blog.google/technology/health/using-symptoms-search-trends-inform-covid-19-research/5.3.3 Enable the Exchange o
250、f Data and Knowledge Among Various PartiesScientific analyses,including experimentation or simu-lation,are among the most important elements in R&D tasks.Both companies and research institutes can be expected to benefit from an interoperable and transferrable ex-change of data via a central collabor
251、ative platform.An industrial data space gives rise to a“network of trusted data”that creates a fully connected data exchange net-work among data owners,data users,brokers,and certi-fication authorities along the full length of the industrial value chain.54 A data platform at a central hub can facili
252、-tate the exchange of information among various depart-ments and allow companies to leverage their assets while concentrating on strategic activities.Merck KGaA,a German health care,life science,and performance materials giant,has developed a data inte-gration platform with the goal of advancing can
253、cer re-search.The platform supports the secure,transparent sharing of data among participating researchers and institutions,enables researchers to arrive at insights more quickly,and safeguards ownership of the data.55 5.3.4 Effectiveness and EfficiencyAn interoperable data-sharing platform availabl
254、e in real time and with possibly lower latency will surely speed up the research process,especially public health care management.Googles predictive analytics in health care illustrates a use case involving the application of data-centric ap-proaches to the support of population health manage-ment.P
255、rior to the COVID-19 pandemic,Google devel-oped a predictive model based on machine learning to analyze the search data people had entered about their symptoms and to predict the spread of seasonal flu;the objective at the time was to enable local medical centers to prepare the infrastructure and pr
256、ovide necessary ser-vices in good time.Today,Google provides a dataset of search trends for researchers studying the link between symptom-related searches and the spread of COVID-19 so as to obtain an earlier and more accurate indication of the reemergence of the virus in various regions.565.4 Recom
257、mendations Finance:(1)Integrate a data due diligence procedure into every commercial or financial due diligence.(2)Shift your focus from the provision of reports to the delivery of added-value finance services such as self-service analytics,spend analytics,cash flow forecasting,and insights into cus
258、tomer product profitability that will enable proactive business guidance.Human Resources:(3)Using data to improve organizational efficiency and effectiveness:Total operational costs,whether for re-cruitment or administrative tasks,can be substantially reduced by the use of process automation,e.g.cha
259、tbots.Page 29Detecon Point of ViewData Centric Company5 Data-centric Corporate Functions(4)Measurements of factors related to working condi-tions such as temperature,humidity,movements,and pulse rates can be used to determine whether the work-ing conditions for personnel are sustainable.(5)Using dat
260、a to close the gap between the current com-petence base and the future workforce map:HR depart-ments carry out strategic workforce planning by aiding employees in the design of a cross-functional career path,in the discovery of opportunities for personal pro-motion,or in the preparation for digital
261、work situations,shifting from an operational to a strategic role for the long-term development of a company.Research and Development:(6)Only a data-centric approach in R&D can redefine the R&D process as a dynamic decision loop that direct-ly analyzes market/customer data for updates of the product/
262、service.(7)By pursuing a data-centric approach,a company or organization can proactively create an innovation eco-system,effectively leverage external resources globally,and readily translate research into innovative,market-ready products and services.By pursuing a data-centric approach,a company or
263、 organization can proactively create an innovation ecosystem.Page 30Detecon Point of ViewData Centric Company5 Data-centric Corporate FunctionsData DemocratizationAt the enterprise level,data centricity unifies the data infrastructure and data management system and migrates or integrates data from l
264、egacy systems into new systems allowing the use of data as required.The single units within an enterprise may have previously managed their data themselves,each one using a system incompatible with all others.If the strategic use of data and the assur-ance of data centricity are to be achieved,data
265、inte-grability,data autonomy,and data democratization are indispensable.The training of employees in the use of available technologies may well be vital in this quest.6.1 Autonomy and Data Democratization in a Data-centric Enterprise6.1.1 Democratization of the Data and the Data StackFor decades,the
266、 handling,analysis,and sharing of data in an organization was the responsibility of the IT depart-ments and theirs alone.The insistence that all analytics pass through the bottleneck of the IT department delays data-driven business decisions,which cannot be made until the data become accessible.Data
267、 democratization is the recognition that data are the source of power and 6 Technical prerequisites for Data-Centricity that,when wielded by the right people,can support organizations in making better,more informed,and more data-driven decisions.The adoption of data democra-tization leads to univers
268、al data accessibility throughout the entire company and to the acceleration of the pro-cess of gaining insights from data(speed to insight).57This democratization process cannot take place unless there is a parallel shift in the technological landscape:entirely new analytics solutions that offer a l
269、ower entry threshold than ever before must emerge.Its realization allows end users(even non-technical users and analysts)to utilize self-service analytic tools,to access historical data quickly,and to integrate,probe,and visualize them with the most suitable tools available.Just one example:graphic
270、drag-and-drop user interfaces for cloud com-puting solutions have become available that permit the simple transformation of data or even the training of simple machine learning modes without having to write a single line of code!57(https:/panoply.io/analytics-stack-guide/data-democratization-getting
271、-started/)58(Reference What Is Data Democratization?Why Is It Important for Institutions?.tr)Figure 10:Simple Visualization of the Elements of Data Democratization58Analytic ToolsData AccessData LiteracyPage 31Detecon Point of ViewData Centric Company6 Technical prerequisites for Data-Centricity59(h
272、ttps:/ employees will be expected to access,manipulate,and interpret huge amounts of data,it is important that they understand what they are dealing with.Misinterpre-tation of data must be prevented.A huge inflow of data may become overwhelming and unmanageable if ade-quate consideration is not give
273、n to the end users under-standing of the data.Important considerations when putting data democratization into actual practice include the ways and means by which employees will access data and the quality of the insights that are being lever-aged.Despite simplifications,new data users will still req
274、uire a certain degree of data literacy if they are to make good use of the data.Even drag-and-drop soft-ware demands skill in its operation,and,above all,skills in the handling and interpretation of the data are vital if valid insights are to be drawn from the analysis.Data literacy helps employees
275、to make data-driven decisions,to interact critically with data,to establish effective data governance,and to make ethical data decisions.59A Forrester report has found that 60 to 73 percent of all data within an enterprise remained unused for analytics as of 2019.One of the reasons stopping companie
276、s short of becoming fully data-driven might be the lack of data skills among their employees,limiting data analytic capacity and preventing full workplace productivity.Three-fourths of the employees felt overwhelmed or out of their depth when working with data.In fact,59 percent of employees reporte
277、d a sense of burn-out when using business intelligence and data analytics tools.This level of unhappiness can be averted if employees are trained in the skills required to deal with data i.e.,data literacy.Since data democratization will continue to be an em-powerment process,the data literacy that
278、is crucial can be achieved through training.There are any number of excellent methods for the de-velopment of data skills among the workforce.First of all,software providers themselves have a keen interest in assuring access to low-cost(often even no-cost)learning tools and materials in a broad rang
279、e of formats,whether texts,videos,or interactive,hands-on guided tasks.Training materials covering multiple levels of skill and tailored to reach a mass audience of end users are available as well.In addition,the learning programs competing to offer affordable access to hundreds of thousands of high
280、-quality,interactive courses in many languages are multi-tude.Businesses committed to data centricity and to the significant improvement of data-driven decision-making should give thought to providing the space their employ-ees require to educate themselves;they might,for instance purchase access to
281、 these platforms for their employees or offer time off from their work responsibili-ties so that they can concentrate on acquiring their new skills without distractions.Investments in employees data skills can also take the form of the organization of elementary and advanced courses on company premi
282、ses.An internal certification course may heighten users awareness of the dangers of misinterpretation and misuse of data.Another valuable solution might be hybrid mentoring programs during which employees progress at their own pace in online training programs while enjoying the support of mentors(co
283、lleagues who are experts in their fields)who supervise and pilot the learning process and provide assistance whenever difficulties or questions arise.Page 32Detecon Point of ViewData Centric Company6 Technical prerequisites for Data-Centricity60(https:/ vs ELT for Data Warehouse:Whats The Best Appro
284、ach?|Software Advice)Finally,the creation of communication channels or forums within the company for the users of a specific software can benefit employees in that they both provide and receive peer support whenever there are issues or concerns.606.1.2 A Paradigm Shift:from ETL to ELTThe process of
285、democratizing data access inevitably entails changes in thinking about the ways data should be handled.In the recent past,a new philosophy of data processing has been emerging,one that reverses the pre-vious order of operations from extract-transform-load to extract-load-transform as this allows bus
286、iness users to utilize and integrate the data stored in the warehouse more directly than in the past.ETL stands for extract,transform,and load.Before we elaborate on this con-cept,let us first explain the idea of data pipelines.A data pipeline is a series of data processing steps.The pipeline begins
287、 with the collection of the data from their source.This is followed by a stage of data processing involving a quantity of data received as input,and the result of each stage is output that becomes the input for the next stage.This procedure continues until the processed data reach their final form a
288、nd destination.Data pipelines allow data to flow from their source(e.g.,a register of incoming calls or a smart device)to a data warehouse or lake and from a data lake to an analytics platform,whether they are transformed(filtered,en-riched,aggregated,merged with other datasets)or where machine lear
289、ning algorithms are running,and from there to a dashboard application or back to a smart device for the improvement of its future performance.Data origi-nating from one source can feed multiple data pipelines.For instance,a product review could generate data to feed a real-time report counting produ
290、ct mentions or a sentiment analysis application.Each of these applica-tions relies on unique data pipelines leading to the de-sired result.As numerous concerns about ETL pipelines have been raised recently,a new paradigm as well as new techno-logical developments and possibilities have emerged.ETL p
291、ipelines have been criticized for several reasons.It is impossible to determine upfront what uses for the analyses,transformations,and raw data in the pipelines might appear in the future.The original idea behind the Figure 11:Business Scenarios for ETL and ELT61ETLELTSource and target databases are
292、 same(e.g.,Oracle source and target databases)Source and target databases are diferent(e.g.,Oracle source and SAP target databases)Data volume is largeData volume is small or moderateData transformations are less complexData transformations are compute-intensiveData is unstructuredData is structured
293、Page 33Detecon Point of ViewData Centric Company6 Technical prerequisites for Data-Centricity62(https:/ of the data may evolve,and it may become neces-sary to restructure the data in deviation from the original concepts.Furthermore,the choice of transformation method for the source data in the ETL p
294、rocess is one of specific decision-makers and may well be final and irre-versible.Perhaps only aggregated data in a daily,weekly,or monthly resolution are available while the analysis re-quires unstructured or fine-grained data in a second-by-second resolutions.There may also be a lack of clarity to
295、 later analysts or data consumers as to who trans-formed the raw data and why they were transformed and how the currently available data differ from the original data.Advancements in the field of cloud-based computation and storage have made a new paradigm possible:the EL(T)approach.EL(T)means that
296、the data are extract-ed(E)from their source in a raw format and stored(L loaded)as-is so that they are available for any type of analyses in the future that might be in the form of a so-called reverse ETL process.Reverse ETL means that the data stored in the original format can be repeatedly re-extr
297、acted and transformed in a different way for every extraction,depending on what is required at the specific time.Since different data users may have differ-ent requirements,ELT supports agile decision-making for analysts and promotes data literacy throughout the entire company.In contrast to ELT,EL(
298、T)decouples the extract-load steps from any optional transformation that may take place.The operational use cases,being unique,might or might not require a transformation process.The analytical use cases might require the transformation of the data at some point.The separation of EL from T allows an
299、alysts to choose whatever type of transformation they wish.6.1.3 The Concept of Data Mesh In 2020,Zhamak Dehghani proposed the concept of data mesh,a new enterprise data architecture frame-work that is expected to bring about revolutionary changes in data-driven decision-making.62The digital transfo
300、rmation of recent decades has led to an explosion in the quantity of generated data.The phe-nomenon has forced the development of increasingly powerful and scalable platforms for the storage and management of data and for their rapid extraction and processing,including real-time analytics.Parallel t
301、o this development,advanced analytical tools capable of handling huge volumes of data have emerged.While it is noteworthy that providers have been creating advanced technologies to accommodate and process data,it has nevertheless been determined that even investments in state-of-the art technologies
302、 do not automatically trans-late into faster data-driven decisions.In the authors opin-ion,this is a consequence of a maladaptive approach to data management and data architecture that assumes the objective of a centralized,monolithic,domain-agnos-tic data storage and management framework administra
303、t-ed by an elite group of highly specialized experts(whose accessibility may be highly limited)with no knowledge of the specific domains where the data originate.The author stresses that the management and process-ing of data by teams isolated from the people generating or consuming the data cannot
304、produce optimal results.The difficulties begin with the generation of source data by teams with advanced knowledge of specific data,but who are frequently not data experts.Whenever data are collected for operational purposes,teams do not neces-sarily feel motivated to care about the datas usefulness
305、 to downstream consumers as long as the envisioned purposed has been achieved.The resulting datasets may Page 34Detecon Point of ViewData Centric Company6 Technical prerequisites for Data-Centricitycontain errors,lack transparency,and be described so inaccurately that downstream users are unable to
306、under-stand them;moreover,the documentation data engineers and analysts require to comprehend the material when they begin working with it may be missing.When end-user teams need reports on which they can base their data-driven decisions,they lack direct access to the data and the requisite data ski
307、lls.They are forced to line up a team of data experts and assign to them the task of mining,processing,and adapting the data as appropriate to satisfy the consumers needs.Yet data experts do not have the knowledge in either the domain of the team that generated the source data or of the target user
308、team.Before they can even begin their actual tasks,they must become familiar with the subject matter and communi-cate extensively with the data creators and consumers before they can understand what types of reports or products they have been asked to create,and inquiries about the details of the da
309、tasets that only the latters creators can answer are inevitable.The consequence:as ideas for new use cases proliferate and the backlog of requests to the data specialist continues to grow,processes are delayed and fast data-driven responses become impossible.These delays can be avoided by moving awa
310、y from the centralized data management framework commonly in use today,namely,the collection of all of the companys Figure 12:Structure of the Data Mesh Architecture at JP Morgan6363(Reference How JPMorgan Chase built a data mesh architecture to drive significant value to enhance their enterprise da
311、ta platform-Dustin Ward)Operational SystemsMove DataProduct-SpecifcData LakesShare DataBusinessProcessesLakeCatalogBlue LakeBlueDataLakeCatalogYellow LakeYellowDataLakeCatalogRed LakeRedDataConsumerCatalogConsuming ApplicationConsuming ApplicationConsuming ApplicationMeshCatalogConsumerCatalogConsum
312、erCatalogdata by a powerful data platform and managed by a small group of highly specialized data experts.Page 35Detecon Point of ViewData Centric Company6 Technical prerequisites for Data-CentricityThe author lists the principles that should be followed to make the new framework more data-centric:1
313、.Domain-oriented,decentralized data ownership and architecture(data are locally owned by the team re-sponsible for their collection and/or consumption)2.Data as a product3.Self-service data infrastructure as a platform4.Federated management of computing resourcesThe type of storage for both raw data
314、 and aggregated data products must be tailored to the needs of the spe-cific consumers.Any data made available to consumers should be registered according to a standard scheme so that anyone requiring the data can also find them(self-service data infrastructure as a platform).The experts on the data
315、-consuming teams then have what they need to perform the required analyses and create the re-quested data products such as reports,presentations,and dashboards.The advantage of this proposal is that data experts have direct access to the team members with domain exper-tise and vice-versa,a situation
316、 in which mutual learning takes place.On the one hand,the data experts acquire domain knowledge and a better understanding of the data they are processing,making it possible for them to be utilized more meaningfully.On the other hand,the non-technical domain experts become familiar with the data and
317、 gain valuable skills by working daily with the data engineers.Decentralization of this type significantly shortens the process from data collection to data-driven decision-making because teams needing data from other teams for their own purposes and the teams with the data communicate directly with
318、 one another;as each team has data experts with knowledge of the domain,mutual communication is further simplified.This must not be understood to mean that decentraliza-tion should involve individual domains using incompati-ble data storage platforms.Such a situation would lead to the creation or pe
319、rsistence of data silos,and the effi-cient flow of essential data to their intended users would be difficult,if not impossible.Modern and scalable platforms for the storage and pro-cessing of enormous quantities of data(cloud solutions that enable the creation of data warehouses and data lakes,e.g.,
320、data lakehouses)should be used,albeit it is necessary to rethink their architecture so that each team has full freedom to manage its own data while simulta-neously assuring ease of accessibility to anyone else who may require the information.The governance of the archi-tecture should be regulated by
321、 role-based access con-trol.Organizations can assign different administrators to different sections of the catalog to decentralize control and management of data assets(federated management of computing resources).This hybrid model of a central-ized catalog under federated control preserves the inde
322、-pendence and agility of the local teams of specific do-mains while ensuring the reuse of the data assets among all teams and enforcing a common security and govern-ance model globally.64The author writes that the data products created by each domain team should be discoverable,addressable,trust-wor
323、thy,possess self-describing semantics and syntax,be interoperable,secure,and governed by global standards and access controls.In other words,the data should be treated as a product that is ready to use and reliable.6564(https:/ 36Detecon Point of ViewData Centric Company6 Technical prerequisites for
324、 Data-Centricity66(https:/ author notes in his summary that the data mesh plat-form is an intentionally designed distributed data architec-ture,under centralized governance and standardization for interoperability,enabled by a shared and harmonized self-serve data infrastructure.666.2 The Infrastruc
325、ture Prerequisites for a Data-centric CompanyData infrastructure serves two purposes at a high level:the storage of the operational data(i.e.,the data that are created and collected in the course of ongoing business such as records of financial transactions,sales opera-tions,customer data,conducted
326、projects)and the utili-zation,transformation,and analysis of these and other collected data in support of business leaders to secure better,data-driven decisions.We have already mentioned that the trend to democra-tization of data access has prompted a shift in the ap-proach to data pipelines away f
327、rom extract-transform-load to extract-load(-transform).In place of top-down decisions mandating the format in which data are stored for future users,data are stored in their original format so that future users are completely free in their choice of the transformations that best suit their needs.Whi
328、le the data warehouses commonly used permit the storage of structured data(usually in tabular form),they have serious limitations and are not suitable for data pipe-lines based on the EL(T)model.When data warehouses are in use,some of the raw data may be irretrievably lost.This happens after the ext
329、raction(E)at the stage when the decision about the method for transformation of the raw data is made(e.g.,their aggregation at specific time intervals(T)and their subsequent loading into the target database(L).Current technology trends both the paradigm shift from ETL to EL(T)and the decentralizatio
330、n of data ownership in line with the data mesh philosophy are better served by so-called data lakehouses,a combination of data warehouses and data lakes.Data lakes are platforms for the storage and subsequent extraction of data in their raw form while data warehouses store structured data.Data lakeh
331、ouses combine the functions of the two plat-forms.They make possible,on the one hand,the saving,storage,and extraction of data in their original form and,on the other hand,the creation of data warehouses(structured,tabular databases)that can result(for exam-ple)from the processing of raw data stored
332、 in the data lakehouse.67In summary,an ideal data infrastructure for modern data-centric solutions should possess the following characteristics:The tangible components of a data-centric organization embracing data democratization are data virtualization software,data federation software,cloud storag
333、e,and self-service applications for non-technical users.This type of infrastructure when combined with a paradigm shift toward democratization of access to data and investments in the development of data skills among employees will not only ensure the transition from the monolithic,centralized data platform paradigm to the decentralized data mesh framework,but will also create optimal conditions f