Meet Your Digital Master

MEET YOUR DIGITAL A guide to transforming master data in the MASTER digital age.

As we journey through Traditionally, master data has focused on opened up new possibilities. Traditionally slow a digital era, traditional establishing a few attributes of a critical and selective master data can now be made Master Data Management business entity extracted from transactional more comprehensive, varied and insights- (MDM) is increasingly systems to create a high-quality version powered to provide a near real-time view of unable to serve emerging of the truth. A lot of time and money has critical business entities. This expanded view business and application been invested into this, primarily for driving not only supports quality business reporting requirements. There is consistency of business reporting. but can actually start supporting transactional digital applications, making real-time decisions a need to expand the This traditional approach has an increasingly by leveraging the “new” master data attributes. realm of master data limited impact in the digital age. The characteristics Master data now, more than ever before, to create value as of data are evolving rapidly and today’s business needs to be highly business outcome focused, business applications needs have become more granular and require adaptable and scalable across the enterprise. get a digital makeover. near real-time responses. Data is not simply confined to structured, internally-generated In this article, we look at how to evolve beyond a transactional data but now spans heterogeneous, traditional MDM to a “Digital Master”, an enterprise machine-generated, unstructured and also that is near real-time, analytically-powered and externally-sourced datasets. directly influences business transactions in a digital world. The end result is business processes The changing needs and evolving data, that consistently deliver tangible outcomes and combined with the advent of analytics, machine impacts, powered by these master entities. learning and cognitive technologies have Meet Your Digital Master 2

Contents 4 Then and Now 24 Exploring the Architectural 10 Shift from Golden Record Components to Golden Profile 28 Digital Master 15 The Contextual and Analytical Powered Business Elements of a Digital Master 32 Deploying the Digital Master 18 Laying the Foundation for 36 Transitioning to the New a Digital Master 39 Conclusion 21 Architecting the Digital Master Meet Your Digital Master 3

Then and Now MDM has traditionally been the framework of processes and technologies aimed at creating and maintaining an authoritative, reliable, sustainable, accurate, timely and secure data environment. Meet Your Digital Master 4

Master Data Master data is the consistent and uniform Increasingly a business requires master core dataset within an enterprise that describes entities to support analytical applications and entities around which business is conducted. not just serve as a reference. This dynamic view Master data includes hierarchical or dimensional of master data requires probabilistic inference data that categorizes core objects of business of data characteristics, which leads to profiles for analytical or reporting purposes. rather than records. This definition, accurate as it is in the context Thirdly, this definition does not refer to of traditional MDM, is limiting in the digital age. the interactions and relationships that exist The “consistent and uniform” designation implies within the data entities across an enterprise. that the data represents the single version of the The focus is on logical and integrated models truth, i.e. a golden record. However, it does not of structured datasets. In the digital age, cover external data sources that provide context datasets now increasingly include unstructured in a dynamic manner. Secondly, it also implies data in batch, real-time and interactive modes that the data is infrequently changed. Is that requiring contextual analytical models. really a positive feature in the digital age? Meet Your Digital Master 5

Challenges Organizations should reconcile investment in traditional MDM to help meet their business needs with digital transformation that makes use of evolving MDM capabilities. Many recent MDM projects have fallen short of delivering on their full potential. There could be several factors behind this, such as: • Implementations can take too long • Creating a data governance framework and cost too much. and measuring and resolving data quality • Project scope is often much too large. issues can be difficult. • Business and IT teams are sometimes poorly • Scaling MDM solutions to deal with the coordinated for achieving the desired end state. volume and complexity of data (especially • Traditional approaches can lag in with increased use of unstructured, keeping up with new technologies. digital data) can be a challenge. • Solutions can delay inclusion of new data sources. Meet Your Digital Master 6

New Digital Demands From our point of view, the back office-driven These business-defined approaches usually reflect the approach runs contrary to the way the business complex, sophisticated needs of many organizations: and its employees need to function if they are to achieve their goals and objectives. • Working from the desired end state and/or from the In Case Study 2 on the following page, the requirements of complex business problems. organization wanted to recognize consumers • Choosing contextual and analytic solutions over traditional as individuals wherever they interacted with MDM tools that can handle new approaches for multi-dimensional the company and connect this identification and complex hierarchical data, including social and commercial with intelligence which could enhance the graphs that underlie digital business use cases. customer experience and increase earnings. • Utilizing new technologies that simplify the development of industry-focused business applications that take advantage The approach needed to be transformational and of master data to help solve business problems. evolutionary. Transformational in the sense that we • Transforming model-driven approaches to be more flexible see more overtly business-defined approaches and agile so they can work with existing business processes that enable the technologies underlying MDM. (instead of creating new processes) and help deliver business Evolutionary because we see the expansion value. IT and business teams can then transition to optimizing of the meaning of MDM in the digital age. various processes for additional value. Meet Your Digital Master 7

Case Study 1: Case Study 2: A life sciences company was severely held back from An organization in the travel industry found it very difficult promoting drugs to individual healthcare organizations. to combine consumer data with visit and hotel information The traditional solution made it difficult to pull together using their traditional solution. They could not recognize critical master data such as: customers without authentication nor anonymous visitors to their websites. Call centers did not have visibility into consumer profiles or prior consumer touchpoints and • Which insurance companies or payers accepted by had limited ability to update customer preferences. the healthcare organization actually cover the drugs. They were limited to a few offers and did not have • And for those payers, which insurance plans provide coverage. the ability to customize based on recent activity. • What other sales teams in the same company were already promoting to the healthcare organization. • Attributes that derive from complex hierarchies and networks of relationships. • What contracts for the pharmaceuticals company were already in place. Only after incorporating changes to enable the Digital Master could the sales team determine if any business opportunities existed before actually having to approach a healthcare target. Meet Your Digital Master 8

The Changing Face of Data In addition to the evolving requirements, the data has changed Case Study 3: significantly and with it the resulting disruption from digital In a recent project, a hospitality company embarked on data on traditional MDM capabilities. If we take a few of these a digital data platform project with Accenture to ensure capabilities and drill down, it quickly becomes clear that that digital master data consumption use cases were traditional methods are not sufficient. met. The digital data platform provided the foundation For example, match and merge for digital data is too complex, to capture, curate, process and store emerging data with a wide variety of digital datasets that run too slowly with types. The objective was to drastically improve customer today’s data volumes. Hierarchical management is too complex, experience, retain loyal customers, prevent churn and especially considering the nature of digital data with its increase market share through personalization and interrelated complex data relationships and formats. truly understanding customers. Furthermore, most enterprise data is either unstructured or Only after incorporating these digital changes was the semi-structured. New techniques are required to adequately organization able to recognize individuals at any touchpoint capture, curate and store these types of data. These data types across channels and enrich their view with intelligent also require different strategies for metadata capture and insights. By using a Digital Master that provides a 360-degree management. Matching and merging these types of data usually view, they have begun the journey to a smarter experience require not only deterministic, but probabilistic and fuzzy matching for their customers, generating increased earnings and algorithms. These probabilistic matching and merging methods enable significant business uplift. the contextual and analytical models necessary for dynamic profiles. Meet Your Digital Master 9

Shift from Golden Record to Golden Profile Examining the MDM landscape reveals a shift from golden records to golden profiles powered by online, transactional and dynamic data characteristics. We could say that there are two fundamental drivers. The first is the new onus placed on master data, and the second is the changing nature of digital data. Meet Your Digital Master 10

Reset Traditional MDM focused on establishing a sanitized single version of the truth. The process usually involved identifying missing or erroneous entries in highly structured datasets, eliminating duplicate records while integrating multiple data sources. While this is still vital, its value alone has diminished in the digital era. The new Digital Master needs to accommodate the evolving requirements and emerging digital data. Rethink The general consensus is that these evolving requirements golden records by validating and eliminating duplicates are becoming more granular and real-time in nature, in and integrating multiple MDM solutions. Changes in data addition to being much more context and insight driven. become rather obvious. There is a shift from traditional golden records to dynamic golden profiles. This also stems from the need to empower We are now producing an unprecedented amount of business users to be able to consume more data and data, including digital data sources, machine-generated, enable patterns that allow for self-service. unstructured, semi-structured and streaming data. This evolution from back office IT-centric golden records These two drivers have given rise to aspirations to modify to business user-centric golden profiles has resulted in or augment traditional MDM to achieve a more complete the need for complementary models. Integration and and dynamic view of data entities. The requirements of logical models (associated with traditional MDM) produce MDM have evolved in the era of digital data and data lakes. Meet Your Digital Master 11

Figure 1: Contrast Between Golden Record and Golden Profile Golden Record Golden Profile Traditional MDM New Digital Master The asks (requirements) The asks (requirements) Identify missing or erroneous entries Expand to context and insights Validate addresses or phone numbers Provide data insights directly to end users Eliminate duplicate records Support real-time solutions with capabilities Cleanse names and addresses Accommodate big data sources Integrate multiple-point MDM solutions Create a 360-degree view from data The solution models The solution models Integration + Logical Models Contextual + Analytical Models Copyright © 2019 Accenture. All rights reserved. Meet Your Digital Master 12

Re-Imagine Contextual and Analytical Models are now Figure 2: Customer Data Entity employed to help accommodate these new drivers and to help evolve and transform the Third-party Traditional Digital Master. These models associated with Information Golden Record the new Digital Master produce a golden profile by providing data insights directly Credit score Name Jane Doe to end users, supporting real-time decision Professional affiliations Address 1 Accenture Road making and leveraging digital data sources Memberships Dallas, TX 00000 Liens Phone 777-777-7777 to create a 360-degree view. So, the question Public records Age 36 then is, what is this golden profile, and how is it different from a golden record? Figure 2 represents a customer data Unstructured Unstructured entry used in a recent digital master data Internal Data External Data implementation that Accenture conducted for an insurance firm. The key objective was Call center conversations Conversations fraud prevention. The company already Point of sale data Reviews had the traditional golden record. Click stream Tweets/Blogs Network data Images and pictures Videos Copyright © 2019 Accenture. All rights reserved. Meet Your Digital Master 13

This customer data entity goes further than traditional data Case Study 4: attributes to provide a meaningful context and a panoramic We recently helped a large regional airline increase its view of critical data entities beyond just attributes (or what you topline sales by expanding the traditional MDM capabilities, get from the golden records). It provides dynamic actionable introducing relevant technologies and models to create insights in real-time from other sources like unstructured a master graph. The dynamic profiles produced for the internal and external data and third-party sources. graph were successfully used to enable real-time inventory These additional data sources are typically not leveraged in tracking and an alerting mechanism for the company traditional MDM implementations. In this implementation, being to make targeted decisions. able to leverage online real-time transactional data, combined with Our experienced perspective introduced a data engineering contextual and analytical models, enables the insurance company pipeline that was fundamental to combining traditional to reduce the percentage of false positives by almost 50 percent models with contextual and analytical ones. This empowered and gain significant savings from fraudulent claims payouts. and enabled business users to consume master data in Golden profiles are the foundation a granular, contextual, analytical and real-time manner. for modernizing revenue management systems in travel-related industries. Meet Your Digital Master 14

The Contextual and Analytical Elements of a Digital Master Requirements and data types have evolved beyond traditional methods, and this new Digital Master essentially combines traditional MDM capabilities with contextual and analytical models to accommodate new requirements in the era of digital data and data lakes. Meet Your Digital Master 15

Contextual Contextual master data models use semantic representation and storage databases to collect and link master data with additional attributes, metadata, transactions and digital data. In contextual master data models, information is blended and personalized to address different functional needs—only the data that is relevant is provided. For example, the relationships between customers and other entities can be extremely hard to decipher in traditional methods, which stand in the way of personalizing customer experiences. These can now be readily and dynamically represented using contextual models and graph databases. Combining these contextual representations with artificial intelligence (AI) and machine learning can enable deep learning capabilities and greater insights. Meet Your Digital Master 16

Analytical Analytical master data models combine master data capabilities with analytics. Figure 3: The “New” Digital Master These models translate master data directly into insight. For example, a customer can be scored daily to calculate a “trust” score —which is then used as an input in fraud-detection models. In addition, analytical models incorporate AI and machine learning algorithms in probabilistic matching and merging. Contextual model Analytical model In a recent master data implementation for identity resolution, we employed Real-time Services Architecture algorithmic matching on heterogeneous data types from real-time streaming logs to create an identity graph for cross-channel devices. The traditional Data Data methods produced deterministic matching with limited opportunities to Standardization Security expand the advertising segments. Data Rather than enforcing format and meaning to facilitate advertising data Data Management Validation exchanges, AI and machine learning enabled us to discover patterns in data, ReM as well as propose associations, correlations and adaptations. To accommodate p as Hierarchy o t Management s er i the added complexity and sophistication of the client’s needs, data entity to Da r Match y t references were better served by the contextual and analytical master data Compatibility a Metadata solution over traditional MDM tools that are built with a relational database at its Management core. Digital master data solutions expand this with analytics, graph databases, machine learning and near real-time analytical visualization. Batch Architecture Copyright © 2019 Accenture. All rights reserved. Meet Your Digital Master 17

Laying the Foundation for a Digital Master Organizations can choose from many different data technology components to build an architecture that supports new digital master data management through an Intelligent Data Foundation (IDF). Accenture believes introducing a data engineering pipeline is fundamental to being able to combine traditional and contextual and analytical methods to create this new digital master data. Meet Your Digital Master 18

Intelligent Data Foundation (IDF) The fundamental step towards building a Digital Master is to modernize your data supply chains to bring in contextual and analytical information built on a solid IDF. Key IDF capabilities should provide the foundation for: • Data Capture staging, ingestion and metadata management. • Data Curation data protection, standardization and promotion. • Data Provisioning business rule and data warehouse. • Data Exploration unification and aggregation. • Data Consumption BI reporting, advanced analytics, machine intelligence and data services. In addition, IDF captures security, governance, data management, monitoring and operations. Meet Your Digital Master 19

The layers are typically built from components Figure 4: Digital Data Supply Chain Architecture that include a big data platform, streaming and complex event processing, ingestion Application framework, in-memory databases, cache Service Interface Layer clusters, data warehouse, NoSQL, analytics, AI, machine learning and search engines. Traditional Relational Distributed Cache In-Mem DB NoSQL IDF can help businesses bridge the gap between traditional MDM and emerging data realities. The focus is to enable sustainable, Query Engine trusted intelligence by empowering businesses Streaming to consume enterprise data widely delivered Event In-Mem Analytics Processing Big Data Core at speed and scale. This enables the Search transformation and evolution of traditional MDM to the new digital master data. So, how do we design the Transfer (Batch, Stream, Interactive) digital master data solution? Sources Copyright © 2019 Accenture. All rights reserved. Meet Your Digital Master 20

Architecting the Digital Master Architecting the new Digital Master implies introducing contextual and analytical models and leveraging (or integrating with) the IDF to provide the expanded master data entity view. Meet Your Digital Master 21

This new digital master data conceptual Figure 5: High-Level Conceptual Digital MDM Architecture architecture consists of common services, Operational Application Analytical Application the IDF, an integration hub and an information hub as shown in Figure 5. Service Interface Layer These hubs represent the three logical data hubs that store master data in various stages of the pipeline process. Logical & Integration Hub Contextual & Analytical Hub In subsequent sections, these logical hubs will also reflect on the deployment options and architecture design options. Digital Data Hub (IDF) Transfer (Batch, Stream, Interactive) Sources Copyright © 2019 Accenture. All rights reserved. Meet Your Digital Master 22

Recall that the integration hub represents the highly deterministic golden record. In the new digital master data, it remains crucial as the sanitized and trusted “reference” data. It is a foundational set of data elements for the entity that increases confidence as digital data sources are introduced, matching and merging to form an expanded view of the data entity. Also recall that the requirements are evolving to be more real-time and granular. The IDF framework and the information hub can enable the polyglot storage and real-time processing and enrichment of data. It also helps ensure that the relevant analytical enrichment, AI and machine learning algorithms can be applied. Finally, contextual representation can provide the dynamic 360-degree view of the data entities. It is then important to drill down to understand what exactly these blocks or layers do. Meet Your Digital Master 23

Exploring the Architectural Components The architectural components for the Digital Master provide the collective capabilities to progressively enrich master records as it flows through the data supply chain pipeline, increasing the data value for greater context and insight. Meet Your Digital Master 24

Figure 6: Logical Reference Architecture Operational Application Analytical Application 1 The integration hub provides an integrated consistent version of the Service Interface Layer truth, leveraging the investments Consolidated Data Web Knowledge Query & Analytics Decision made in traditional MDM. Views Virtualization Services/API Graph Services Support 1 2 The digital data hub allows for master Semantic y t Integrated Views Information Graph Representation t n data, data analytics enrichment and e i h g l a Hierarchical tc u hme decision support, along with key a c Management i M Mera Qr 3 & t n components for the IDF. a E t Data Privacy Integrated Hub Master Information Hub D & n a me 3 The information hub provides ate D ag the capability to uncover complex Real-time Processing Predictive Scoring an M relationships that exist within the Smart Data Preparation 2 Digital Data Hub (IDF) Machine Learning enriched master data with an & Analytics Enrichment Real-time Query Insights Discovery expanded entity view. Transfer (Batch, Stream, Interactive) Sources Common Services Digital Data Hub (IDF) Logical & Integration Hub Contextual & Analytical Hub Data Movement Data Processing Data Interaction Copyright © 2019 Accenture. All rights reserved. Meet Your Digital Master 25

Logical and Integration Hub This provides traditional data It supports multi-domain master data in Figure 7: Logical and Integration Hub integration and standards to uniformly a single data model. Organizations usually maintain and govern hierarchies and rely on it for the golden records of the Integrated Views data maps. Most organizations may relevant data entities. The database Hierarchical Management already have this in some form storage is typically relational and supports Integrated Hub of MDM solution. deterministic matching and merging. Data Privacy Copyright © 2019 Accenture. All rights reserved. Contextual and Analytical Hub The master information hub maintains The knowledge graph can be seen Figure 8: Contextual and Analytical Hub the semantic representations of master as a snapshot of the entire information interconnected data and provides a data graph and can be based on a particular Information Graph Semantic Representation store for contextual data. The information business use case or query pattern. graph is the semantic representation Master Information Hub that produces the knowledge graph and contextual model. Copyright © 2019 Accenture. All rights reserved. Meet Your Digital Master 26

Digital Data Hub The Digital Data Hub is a central data repository for raw and derived enterprise data. It may typically be an enterprise data lake. It includes real-time and batch-processing capabilities. IDF, as earlier stated, consists of the end-to-end data supply chain pipeline that enables this new digital master data. • Real-time processing Real-time detection • Predictive scoring Predictive models are Figure 9: Digital Data Hub of patterns in the data and event triggering are used to score data, delivering precise insights enabled, which in turn promotes prompt decision that facilitate informed decision-making Real-time Processing -making, enhancing the data lake’s processing based on quantitative logic and insights. Smart Data Preparation & Analytics Enrichment capability, insight generation and analytics. • Machine learning Machine learning leverages • Smart data preparation & analytics enrichment IDF processing capabilities to rapidly explore Real-time Query Data preparation is an iterative process for different permutations and generate the best exploring and transforming raw data into model for entity resolution and fuzzy data forms suitable for analytics enrichment, data matching. Machine learning is also leveraged Digital Data Hub (IDF) science experiments, data discovery and for intelligent data capture and curation, analytics. Analytics enrichment infuses digital profiling and data tagging to enhance Predictive Scoring data transactional information with attributes metadata management and data quality. that enable a 360-degree view of data. • Insights discovery Statistical and mathematical Machine Learning • Real-time query Real-time query capabilities methods are leveraged in order to provide both Insights Discovery facilitate insights discovery, data visualization descriptive and predictive analysis of data to and real-time decision support. facilitate query, analytics and decision support. Copyright © 2019 Accenture. All rights reserved. Meet Your Digital Master 27

Digital Master Powered Business To illustrate how this all fits together to enable both user and business applications to leverage digital master data to make timely decisions, consider the industry example on data flow on the following page. This new digital master data shifts the focus from merely powering downstream systems to allowing the business user to actively consume master data in a dynamic, granular, contextual and real-time manner. Meet Your Digital Master 28

A rethinking of what master data means has given rise to opportunities for a more business and outcome-oriented focus. The new Digital Master has challenged our insistence on strict, deterministic single versions of the truth. By expanding what digital data means, we are also seeing an acceptance of probabilistic approaches that aim to increase the number of potential outcomes. These outcomes provide a rich input to AI and machine learning applications, data services and analytical applications Challenge Digital Master Flow We recently helped an organization with the The emphasis in this example and indeed with the new objective of building a cross-device identity graph digital master data approach, is the empowering of the for advertising purposes, which consisted of data business user and applications to consume master data on households, devices and individuals, enriched in a granular, contextual, analytical and real-time manner. with online digital data sources. Services such as bring-your-own-data and broad activation Notice that in Figure 10 on the following page, channels were key requirements. the master data is progressively enriched as it flows through the information pipeline and increases the master data value. Meanwhile, master data along the pipeline enhances the data confidence while helping to convert information gleaned from digital data sources into actionable insights. Meet Your Digital Master 29

Figure 10: Next-Generation Work/Data Flow Service Interface Layer 10 Consolidated Views Data Virtualization Web Services/API Knowledge Graph Query & Analytics Services Decision Support y t Integrated Views Information Graph 9 Semantic Representation t n e i h g l a tc u hme Hierarchical Management a c i M Mera Qr 2 Integrated Hub 7 Master Information Hub 8 & t n a E t Data Privacy D & n a me ate D ag Real-time Processing Predictive Scoring an M Smart Data Preparation 2 5 Digital Data Hub (IDF) 6 3 Machine Learning & Analytics Enrichment Real-time Query Insights Discovery Transfer (Batch, Stream, Interactive) 1 4 Sources Common Services Digital Data Hub (IDF) Logical & Integration Hub Contextual & Analytical Hub Data Movement Data Processing Data Interaction Copyright © 2019 Accenture. All rights reserved. Meet Your Digital Master 30

1 Capture deterministic and probabilistic 4 Augment this linkage with third 7 Merge multiple ID graphs together IDs using Personally Identifiable Information party and external linkage solutions. in ID translation refinery (steps include (PII), device IDs, location, cookies, Here bring-your-own-data is enabled decryption, match/cluster, translate, Universally Unique Identifiers (UUIDs), and combined with analytical create combined graph). The information derived behaviors/patterns. These can enrichment to add enhanced hub enables the relationship linkage. come from internal, third party and external profile data elements to an entity. online data sources. Intelligent tagging is 8 Create persistent IDs for each person key to breaking out behavioral patterns. 5 Create or update the ID graph by based on the graph, i.e. global or merging online and offline graphs ultimate ID. This ID will be tied to each 2 Develop a “device fingerprint”—this is together. The digital hub is the repository device the person is associated about identifiers stored in identity graphs for all the graphs in a database format. with according to the graph. (including network information, device This hub has the polyglot nature to IDs, browsing behavior or third party maintain all the data elements that then 9 Build the complete digital identity cookies) to create a “fingerprint” that can form the basis for the profile graph. resolution data universe. predict someone’s identity or household Build various activation services on with a fair degree of accuracy. 6 Simulate real-time graph updates. 10 The digital hub also has the real-time the service interface layer for business 3 Apply deterministic and probabilistic query and updated capabilities to ensure user interaction and ID resolution matching to create linkage from device the ID graph is refreshed and up to date. data consumption. to person. Machine learning is leveraged here to perform probabilistic matching. Meet Your Digital Master 31

Deploying the Digital Master Having discussed the functional views and logical components, how do we translate these to actual solutions or implementations? How do we make informed decisions around the best deployment approach? Meet Your Digital Master 32

Implementation Patterns Fundamentally, there are two implementation patterns—single container and multiple containers. The differentiation between these two patterns is based on the “container” each of the logical hubs is implemented in. Single Container In the single container approach, all three Figure 11: Single Container Table 1: Deployment Strategy logical hubs are contained within the big data platform or data lake. It also ensures that Single Container Green Field on Data Lake the functionalities of each of these hubs are Cloud or on-premise solution implemented within the data lake. Only a sub set of core MDM implemented From an on-premise standpoint, typical tools Logical & Contextual & Integration Hub Analytical Hub Graph-based master and integration could include Hadoop/Spark/HBase for the Digital Data Hub, Hive or Kudu for the integration data repository hub and Neo4j for the information hub. S3, EMR, Digital data technologies— Snowflake, DynamoDB or Neptune could form Hadoop, Cloud, NoSQL (graph, columnar) the basis for a cloud single container approach. Digital Data Hub (IDF) Flexible integration with existing enterprise MDM solutions Copyright © 2019 Accenture. All rights reserved. Meet Your Digital Master 33

Multiple Containers In the multiple container pattern, one or more Figure 12.1: Multiple Container, Type 1 Figure 12.2: Multiple Container, Type 2 of the logical hubs can be in different containers. The integration hub may be in a separate Container One Container Two Container One Container Two traditional master data container. Many clients have made significant investments in traditional MDM and would like to leverage this as much Logical & Contextual & Logical & Contextual & as possible. Tools associated with the traditional Integration Hub Analytical Hub Integration Hub Analytical Hub MDM container will include Profisee, Informatica, IBM InfoSphere, Teradata, Oracle, Talend, SAS and TIBCO. Container Three Many clients aspiring to go for the single Digital Data container approach may initially have started Hub (IDF) with the multi-container approach. It provides a soft landing and is a more cautious way Digital Data Hub (IDF) of reaching the target state. Copyright © 2019 Accenture. All rights reserved. Copyright © 2019 Accenture. All rights reserved. Meet Your Digital Master 34

In addition to the options for the Table 2: Deployment Strategy separate containers, there are two possible options where each of Managed Services On-Premise Appliance these hubs may be deployed: • Managed Services Usually cloud-based or externally Well-established traditional MDM hosted solution solution providers • On-Premise Appliance As data lakes become commonplace Graph-based master data repository Typically relational database solutions in the enterprise and technologies APIs provided for data ingestion Tight coupling to vendor technology continue to mature, master data and and service delivery stack and appliance recommendation digital data will coexist much more naturally and data will be entirely Big data technologies—NoSQL (graph, columnar) Mature support from vendor mastered in the data lake itself. Example technologies—Reltio, Pitney Bowes Example technologies—IBM, Informatica Copyright © 2019 Accenture. All rights reserved. Meet Your Digital Master 35

Transitioning to the New Most organizations have some form of traditional MDM implementation and are looking to enter the new realm of master data in the digital age. We are observing two transition patterns, which we call “Extend” and “Replace”. Meet Your Digital Master 36

These transition patterns typically involve a two- However, reliance on legacy systems, cost step process, i.e. starting with an existing MDM of maintaining such systems and the data implementation, integrating with a data lake and synchronization requirements needed to eventually migrating all hubs and capabilities to enable the successful transition to the new the data lake. Some organizations bite the bullet Digital Master are some issues to contend with. and go for this end state directly. In the “Extend” strategy, clients opt to keep the traditional master With a “Replace” strategy, clients have an in the existing traditional integration hub while opportunity to build a robust new Digital extending or expanding the capabilities beyond Master entirely around a big data lake. This is a the traditional ones by integrating with a data lake. significant emerging trend, especially with data lakes becoming more mainstream in the digital The information hub is typically decoupled era. Clients can leverage existing data lakes with from the integration hub, which helps in scaling the added flexibility and freedom of building and maintenance, but also in complexity and adaptive intelligent master data models. operations. The reduced time frames for the project build and deployment are an added Other pros of this approach include a more incentive to achieving a minimum viable product seamless coupling of the various logical hubs as quickly as possible. Table 3 on the following and reuse of common services. The cost of page shows the pros and cons of these building from scratch, time to deploy and data transition strategies. lake immaturity are all possible challenges. Meet Your Digital Master 37

Table 3: Transition Patterns Transition Description Pros Cons Extend Integrate existing MDM with big • Leverage existing MDM systems • Legacy MDM data platform (data lake)—data lake • Leverage existing data lake • Data synchronization needed consumes traditional MDM data to enable single source of truth • Information hub is decoupled from integration hub • Reduced project build and deployment time frame Replace Implement digital MDM requirements • Leverage existing data lake • Built from scratch entirely within data lake • Flexibility and freedom in • Time to deploy and cost designing master data models • Data lake immaturity for core • Tighter coupling between traditional MDM functions integration and information hubs • Green field implementation Copyright © 2019 Accenture. All rights reserved. Meet Your Digital Master 38

Conclusion To gain a competitive advantage from its enterprise data, an organization should be able to generate business insights from it. The barrier to entry is at an all-time low, thanks to the advent of digital data. Meet Your Digital Master 39

As we enter this new era of digital data, they have selected. Additionally, they can traditional MDM methods are no longer derive a more comprehensive understanding sufficient. There are two main evolutionary of the truth based on a host of relevant drivers forcing a renewed look at what master external factors. data means to the enterprise—new demands By leveraging our Intelligent Data Foundation, that are placed upon the master data and transforming the role of master data and an expanding universe of data sources empowering the business user to consume that need to be integrated. “online master data” to drive decisions, Traditional solutions and methods are this new digital master data can be characterized by difficult and expensive ready for the digital age. implementations that are complex to maintain. This point of view provides an overview Most importantly, these solutions have very for organizations to use as a starting point little impact in today’s digital world. However, to understand this evolving landscape and by augmenting existing master data with digital familiarize themselves with the appropriate data, organizations are now able to greatly architectural options. expand the view of whatever domain Meet Your Digital Master 40

Next Steps Here are the key actions to help build your organization’s new Digital Master: 1. Data inventory 5. Models, AI and machine learning Start with your most frequently accessed and Identify the right contextual and analytical models time-relevant data. This can allow your initial to cover the business use cases. sources to establish a version of the truth. 6. Consider external data sources 2. Data hubs Look outside your organization for external data Identify current data repositories to understand sources that can be incorporated to complement current needs across the business and the existing reference data and help generate extent to which these data hubs meet the more complete insights through contextual business requirements. and analytical models. 3. Use cases 7. Choose the data technology stack Examine the business use cases and align these Research the optimal deployment methods use cases to existing solutions. Identify the for your organization. potential opportunities. 4. Prioritize individual Digital Master use cases Prioritizing helps you develop a roadmap for implementing the new Digital Master at scale. Meet Your Digital Master 41

Contact About Accenture Applied Intelligence Ekpe Okorafor, PhD Accenture Applied Intelligence, part of Accenture Digital, applies AI and human Senior Principal, Accenture Digital – ingenuity at the core of business to help clients solve their most complex business Applied Intelligence problems. By deploying AI responsibly and combining it with our deep industry [email protected] and analytics expertise, we enable the digital transformation of organizations, Atish Ray extend human capabilities, and make intelligent products and services a reality. Managing Director, Accenture Digital – Follow @AccentureAI and visit accenture.com/appliedintelligence. Applied Intelligence About Accenture [email protected] Accenture is a leading global professional services company, providing a broad range of services and solutions in strategy, consulting, digital, technology and operations. Combining unmatched experience and specialized skills across more than 40 industries and all business functions—underpinned by the world’s largest delivery network—Accenture works at the intersection of business and technology to help clients improve their performance and create sustainable value for their stakeholders. With approximately 469,000 people serving clients in more than 120 countries, Accenture drives innovation to improve the way the world works and lives. Visit us at www.accenture.com This document is produced by consultants at Accenture as general guidance. It is not intended to provide specific advice on your circumstances. If you require advice or further Copyright © 2019 Accenture. details on any matters referred to, please contact your Accenture representative. All rights reserved. This document makes descriptive reference to trademarks that may be owned by others. Accenture, its logo, and The use of such trademarks herein is not an assertion of ownership of such trademarks High performance. Delivered. by Accenture and is not intended to represent or imply the existence of an association are trademarks of Accenture. between Accenture and the lawful owners of such trademarks.