To Become A Data-Driven Enterprise, Data Democratization Is . - Cognizant

Cognizant 20-20 InsightsTo Become a Data-DrivenEnterprise, Data Democratizationis EssentialTo optimise enterprise knowledge, organizations need a modern platform thatenables data to be more easily shared, interpreted and capitalized on by internaldecision makers and by business partners across the extended value chain.Executive SummaryFrom the time the term ’data is the new oil’ originatedin 2006, interest across enterprises to become datadriven has skyrocketed. To become a true data-drivenenterprise, organizations must activate the troves of dataon which they sit to glean insights and foresights that driveinnovation, competitive performance and better customerexperience. This process is called data monetization.Though the term involves money, it does not always haveto be about selling data or insights. There are two waysenterprises can monetize data:December 2020 Internal monetization, which is about unlockingthe value of the data through innovative products,operational efficiency or better customer experience. External monetization, which is about making dataor insights available to partners or other externalconsumers for a price (as information servicesproviders D&B, AC Nielsen, Experian and others do),or making data or insights freely available so thatconsumers can use it to build products (governmententities and educational institutions often do this).

Cognizant 20-20 InsightsIn our view, data monetization will come automaticallythrough the democratization of data, anddemocratizing data requires a modern data platform.A modern data platform helps move data from silosto a common platform, which provides the abilityto bring massive quantities of bits and bytes topower inductive analytics. The platform also reducesthe time required to extract insights from data,which accelerates decision-making. By bringingmassive amounts of data and breaking data silos,organizations can establish a single version of truthand improve the discovery of data assets. Thisdiscovery is critical to provide data democratization.Data democratization is the process of enablingaccess and availability of information across linesof businesses to drive innovation via self-servicebusiness intelligence (BI) and predictive analyticsplatforms or by applying deep-dive data science. It’sa departure from the traditional process, in whichdata is typically owned by a central IT team and thelines of business — that is, decision makers — mustwork through IT to access the data they need.In this paper, we lay out our view on the ways andmeans of creating an insights marketplace thatmodernizes and frees data from its shackles andprovides a way forward for organizations seekingnew and vital ways to innovate and profit fromexisting data stores. We conclude with examplesof how democratization is generating results (seeQuick Take, page 9).2 / To Become a Data-Driven Enterprise, Data Democratization is Essential

Cognizant 20-20 InsightsOvercoming Data Democratization ChallengesBuilding applications directly on modern cloudbased data platforms comes easily for the digitallynative organizations that have emerged during thismillennium. The advantage is clear: a single, unifieddata store that is accessible to a wide community ofconsumers instantaneously. However, this approachrepresents a huge leap for established enterprises(see Figure 1).One goal of data modernization is the creation of asingle trusted data platform that can unify existingsilos, making the combined and standardised dataavailable to a wide community with clear governanceand security controls in place.With data modernization, organizations createa cloud-enabled ecosystem that brings togetherdata from across the enterprise. This helps teamscross-pollinate data in ways that uncover actionableinsights and, importantly, makes it available forconsumption on a real-time basis.A modern data platform consists of three maincapabilities: A responsive data architecture that is extensibleto changing business and market needs —meaning it can process a range of the four Vs ofdata (volume, velocity, variety and veracity). Intelligent data management, which includesproper governance mechanisms and metadatamanagement. This provides the right level ofcontrols on the data and renders the data trustable. Delivery at scale, which encompasses automationand DevOps methods needed to truly deliver at scale.Additionally, the modern data platform is set up withintelligent management capabilities that enable democratization and monetization of data (see Figure 2). Data platform: This provides a foundation fordefining a new modern data platform or forextending a legacy environment to the cloud, whereit can ingest data from a variety of systems of recordwithin the enterprise and at the desired frequency ofbatch, real-time, streaming — or through applicationprogramming interfaces (APIs). External/internal data: This enables theingestion of relevant data wherever it mayreside into the platform via any ingestionmethod available (through feeds, for example, ortechnology-enabled data exchanges).Typical Challenges Inhibiting CollaborationData and analytics shared-services organizations face myriad challenges:LegacyArchitectureDataPlatformvs. InsightsPlatformRedundant dataInability toand analyticalgauge usageassets acrossand ROI of datamultipleand analyticalsystemsassetsFigure 1Unreliableand MultipleVersions ePoorSlow turnaroundLack ofInward looking,performance ofof requestsself-service andbottom upsystems due todue to ads to ‘shadowon data, noand dataand reuseIT’ teamsperspective onproliferationamong businessmaximization/users, datamonetizationscientistsof data assetsinternally orexternally3 / To Become a Data-Driven Enterprise, Data Democratization is EssentialUnorganizedKPIs andAssetsInsights spreadacross BI tools,models and data,complicatingseamless access.Users end upsearching forKPIs

Cognizant 20-20 Insights Data governance: This enables the followingcritical features: Data quality, which helps define and monitorthe quality of the data assets. Data catalogue, in which all data assetsare made available along with metadataand lineage. The data assets also provideinformation on the quality service levelagreements (SLAs), thus making the datamore trustable. The metadata available in thedata assets will also include currency stamps(“use by”). Master data, which helps create a single sourceof truth for shared master data assets. Thisreduces data duplication and increases thequality of data. Data security, which helps establish a rolebased access control (RBAC) mechanism togovern data access, which must be authorisedby the data owner before access is provided tothe user. It also provides the audit of the userswho have/had access to the data, along withtheir level of access. Data compliance, which helps implement datacompliance requirements based on the typeof data. Data catalogue: This provides a repository ofthe information available, its lineage, and qualitymetrics on the data products to enable datademocratization and self-service. Data products: Data engineering principles mustbe applied to create data products that enableseamless browsing and consumption. BI assets: Metadata and tags for the BI and Analyticsassets must be defined and built by users; they canthen be re-used across the enterprise. Machine learning (ML) assets. Metadata and tagsfor re-usable ML models and related feature setsare defined and built by data science communitiesacross the enterprise for search and subscription. Insight marketplace: Democratization is enabledthrough a marketplace interface with which userslook for the data assets they need, and whichenables subscription through workflows.The typical data democratization value chain isdepicted in Figure 3. The goal is to consolidatedata into a modern platform on which internal andexternal data is made available for business users,data scientists, partners, and external consumers.Each data asset or feature made available toconsumers should be in the form of a data productthat can be consumed in a self-service model.Insights Marketplace Needs a Dedicated PlatformEmbed and integrateexternal sources into datalandscapeSet up adedicatedmodern dataplatform orextend existingplatformStore, cleanse andprepare data forinsights generationLeverage analytics and data science bybuilding KPIs, reports and dashboardsPerform data science and analytics to address usecases in service quality, forecasting and preventivemaintenanceImplement state-of-the-artaccess control, security andgovernance mechanismsFigure 24 / To Become a Data-Driven Enterprise, Data Democratization is EssentialEnable democratizationthrough the insightsmarketplace

Cognizant 20-20 InsightsThe Insights Sharing Value ChainData ProducersDataAcquisitionData AggregatorsRAWDATADataManagement,Transformation &OperationsData ConsumersTRANSFORMED,CLEANSED, STD.& ENRICHEDDATAMI & InsightsBenefitsto CoreBusinessML modelsDataGenerationData &InformationProductsNewProducts& RevenueOpportunitiesData GovernanceData Persistence & Infrastructure (HW & SW)Figure 3Creating an Insight MarketplaceOnline retailers such as Amazon and eBay haveperfected the model of self-service in the world ofshopping. A typical online marketplace for productsworks as shown in Figure 4.Multi-Vendor Marketplace StructureMarketplaceStore oductsFigure 45 / To Become a Data-Driven Enterprise, Data Democratization is EssentialCustomers

Cognizant 20-20 InsightsThis model can be applied directly to a data-drivenenterprise to enable self-service of data products, asillustrated in Figure 5.An insights marketplace offers a nice way todemocratize the data within an organization in itsjourney toward becoming a data-driven enterprise.An insights marketplace is an interface whereby usersacross the enterprise can search for data products,BI, analytics and ML models that have been producedacross the organization. The insights marketplacealso makes the process of getting access to the dataseamless by automating many data governanceactivities, such as data security and provisioning.Users will be able to request data assets that are notavailable in the marketplace, as well. The insightsmarketplace can drive the following (see Figure 6)(next page): Collaboration across teams and business units. A reduction of duplicated efforts to ingest dataand create data products. Secured data democratization. Internal monetization of data.The Data-Driven EnterpriseInternal Data Sources CRMFinanceHRERPCustomer ServicesExternal Data Sources WeatherSocial MediaCredit ScoringEtc. DataPlatformBI/ReportsRaw DataFeaturesAI/ML modelsInsightsMarketplaceData Governance, Security, eCommerceFigure 56 / To Become a Data-Driven Enterprise, Data Democratization is alconsumers