WIXLIB

Achieving Business Value throughBig Data AnalyticsPhilip RussomTDWI Research Director for Data ManagementOctober 3, 2012

Sponsor2

SpeakersPhilip RussomResearch Director,Data Management,TDWIBrian NgDirector, Enterprise Services,HP

Today’s Agenda The Need for Business Valuefrom Big Data Definitions of Big Data Analytics Use Cases for Big Data Analyticsthat deliver Business Value The Future & How to Prepare for It

Background The quantity and diversity of Big Data has been exploding for years– Traditional applications grow larger & more numerous every day– Older big data sources: RFID, call detail record, machine/robotic data– New big data sources: sensors, social media, new Web apps User organizations are starting to achieve business value from big data– The consensus today is that Advanced Analytics yields valuable business insights– As long as big data is managed well and treated to the right forms of analytics Today we’ll look at how Big Data Analytics can deliver business valueIn your organization is big data considered mostly a problem or mostly an opportunity?70%30%Opportunity – because it yields detailedanalytics for business advantageProblem – because it's hard to managefrom a technical viewpointSource TDWI. Survey of 325 respondents, June 2011

Definition of Big Data Analytics It’s where advanced analytic techniques operate on big data sets. It’s about two things: big data AND advanced analytics.Big Data isn’t new.Advanced Analytics isn’t new.– Their successful combination is new.– Both users and technologies arenow more capable of success. The combo is new & technical.– But hasn’t yet aligned with business.BigData Big DataAnalyticsAdvancedAnalytics– The two have teamed up to leverage big data.– The combo turns big data into an opportunity.

The “3 Vs” of Big Data summarizetechnical propertiesBusiness Value should be the 4th V,since this is what IT must BUSINESSVALUE

Defining Advanced AnalyticsOLAP & its Variants Users have this They’ll keep & grow it OLAP won’t go away–Online Analytic Processing (OLAP)––––Extreme SQL––Advanced Analytics Discovery oriented Excels with Big Data Experiencing strongadoption by users–Uses data mining and/or statisticsto anticipate future events.Multi-Structured Data Analytics–––Uses well-known SQL-based tools & techniques.Relies on long, complex SQL statements.Predictive Analytics––It’s somewhat rudimentary, but required.Demands multidimensional data modeling,but works well with most EDWs.There are multiple approaches to OLAP.Natural language processing (NLP)Search, text analytics, sentiment & social analytic appsOther Analytic Methods––Visualization, artificial intelligenceAnalytic database functions: in-database analytics, inmemory databases, columnar data stores, appliances, etc.

TDWI SURVEY SAYS:Opportunities for Big Data Analytics Anything involving customers benefits from big data analytics– better-targeted social-influencer marketing (61%)– customer-base segmentation (41%)– recognition of sales/market opportunities (38%) BI, in general, benefits from big data analytics–––– more numerous and accurate business insights (45%)understanding business change (30%)better planning and forecasting (29%)identification of root causes of cost (29%)Specific analytics applications are likely beneficiaries– detection of fraud (33%), quantification of risks (30%)– market sentiment trending (30%)Source TDWI. Survey of 325 respondents, June 2011

USE CASEExploratory Analytics with Big Data Big Data enables exploratory analytics. Discover patterns and new facts thebusiness didn’t know––––Customer base segmentsCustomer behaviors and their meaningForms of churn and their root causesRelationships among customers andproducts– Root causes for bottom line costs– State of biz today; predict future events

USE CASEAnalyze Big Data You’ve Hoarded Yes, it’s true:– Many firms have “squirreled away” largedatasets, because they sensed businessvalue, yet didn’t know how to get value outof big data. Finally understand:– Web site visitor behavior– Products of affinity based on eCommerceshopping carts– Product and supply quality based onrobotic & QA data from manufacturing– Product movement via RFID in retail

USE CASEBig Data Analytics per Industry The type and content of big data canvary by industry, thus have differentvalue propositions per industry:– Call detail records (CDRs) intelecommunications– RFID in retail, manufacturing, andother product-oriented industries– Sensor data from robots inmanufacturing, especially automotiveand consumer electronics

USE CASEAnalytics for Unstructured Big Data Tools based on natural languageprocessing, search, and text analytics(plus new platforms like Hadoop) providevisibility into text-laden businessprocesses:– Claims process in insurance– Medical records in healthcare– Call center and help desk applications inany industry– Sentiment analysis in customer-orientedbusinesses, with both enterprise andsocial media big data

“I love/hateyour product!” USE CASECustomer Analytics withSocial Media DataCustomers can influence each other by commenting on brands, reviewingproducts, reacting to marketing campaigns, and revealing shared interestsPredictive analytics to discover patterns, anticipate product/service issuesMeasuring share of voice and brand reputationBroader input for customer satisfactionUnderstanding sentiment driversVoice of the customer analyticsDetermining marketing effectivenessIdentifying new customer segments

USE CASEBig Data forComplete Customer Views Big data can add more granulardetail to analytic datasets:– Data from all customer touch points– Broaden 360-degree views ofcustomers and other entities, fromhundreds of attributes to thousands– For more detailed and accuratecustomer base segmentation, directmarketing, and other customeranalytics

USE CASEBig Data Can Improve Older Analytics Big data enlarges and improves data samples forolder analytic applications:– Any analytic technologies thatdepend on large samples,such as statistics or data mining– Fraud detection– Risk management– Actuarial calculations

USE CASEAnalytics with Streaming Big Data Monitoring & Analysis in True Real Time– Energy utility, communication network;any grid, service, facility– Surveillance, cyber security,situational awareness– Fraud detection, risk calc– Logistics, truck/rail freight,mobile asset mgt Near Time– Review of loan applicationssubmitted online

A LookInto the Future ofBig DataAnalytics Big data analytics is here to stay– It will spread into more apps in moreindustries, becoming mainstream Big data will be less of a problem– Due to advances in storage, clouds, CPUs,memory, databases, analytic tools, etc. Analytics will draw biz value from big data– That’s why the two have come together New types of analytic apps will appear– Old ones will be revamped110100110101110100100101011 Big Data Analytics is mostly batch today– Will go real time as users/techs mature Analytics is new competency for many shops– They will hire & train, plus acquire tools andseek professional services

Recommendations Insist on business value from big data–– Go beyond reporting and OLAPinto advanced analytics– You need “discovery” analytics,but reporting and OLAP won’t go awayEmbrace the brave new world of big data– Don’t merely hoard it in a cost center thatwastes valuable storage & other resourcesThe path to business value is through analyticsData from Web, machine, and social sourcesUpgrade, extend or distribute your BI/DW tech stackand other software portfolios with technologies for big data and analytics– Change is needed to accommodate analytics with big dataGive the business what it needs– Discovery analytics to understand change, find opportunities– Broader, more complete views of customers & other business entities– Analytics tailored to your industry and your organization’s unique requirements

Achieving businessvalue through bigdata analyticsHP Enterprise Services, Information Management andAnalytics,Brian Ng / Oct 2012 Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice.

Our point of viewThriving in the age of big sFraudWe are at a fundamental inflection point in the evolutionof information and intelligence.Traditional approaches, architectures and organizationsmodels were not designed for today’s complexity.Leadership will be defined by those who excel in informationsciences, via innovative solutions, advanced technologies & newtalent models.21 Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice.

Use cases and architecture Copyright Copyright2012 opmentCompany,Company,L.P. L.P.The informationThe informationcontainedcontainedhereinhereinis subjectis subjectto changeto changewithoutwithoutnotice.notice.

1. Unstructured and structured analysisLogical architectureService managementPortfolio managementOperations managementEvent processingComplexeventprocessingRules engineSOA servicesHumanData acquisitionLanguageSentiment,Mark-up &Relational DBMS Integrate DataData captureGovernance23data warehouseData transformationInternal dataUnstructured dataRulesgeneratorVisualizationRepositoryRich MediaSemistructureddataExternal dataApplicationsRaw DataRepositoryCaptureStructureddataAnalysis and reportingApplicationsAnalyticalData miningData MartengineMaster data mgtMatch andcombineIntegrationAggregationPopulate riseDWDBMS(content mgtsystems)Data governanceOLAP reportsVirtualizationSQL analyticsengineMaster dataDerive metadataandDataindexqualityVisualizationStatic andAnalysisData mart(e.g. OLAPDatacube)ReportingSearchDashboardsengine OlapmartsNoSQLData mining(e.g.VisualizationMapReduce)engineData audit, balance and control Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice.StatisticalanalysisStatistical sand alertsPredictiveanalysis

2. Machine generated data streamsLogical architectureService managementPortfolio managementEvent processingApplicationsComplexeventprocessingRules engineComplex EventProcessingRulesEngineSOAservicesData acquisitionSensorNetworkData captureInternal dataUnstructured dataGovernance24ApplicationsRulesgeneratorMarkup, stream,integrateRelational DBMSdata warehouseData transformationSemistructureddataExternal dataAnalysis and reportingRepositoryRaw DataRepositoryCaptureStructureddataOperations managementDataVirtualizationAnalyticalData miningData MartengineDerive metadataandDataindexqualityMaster data mgtMatch andcombineIntegrationAggregationPopulate riseDWDBMS(content mgtsystems)Data governanceData mart(e.g. OLAPDatacube)ReportingSearchDashboardsengine OlapmartsNoSQLData mining(e.g.VisualizationMapReduce)engineData audit, balance and control Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice.Dashboardsand alertsStatisticalanalysisStatistical analysisOlapReal-timecubesanalyticalRDBMSStatic andRealOLAPtimereportsvisualizationand analysisSQL analyticsengineMaster DBMSVisualizationModel and Ruledevelopment.Predictiveanalysis

Use case: Insurance claim fraudFraud detectionBusiness issue Insurance claim fraud continues to be a major costBig Data sources Claims form (human language) Contact records (call center logs, audio, email,instant message, video calls)Process Sentiment analysis and meaning-based scoring Input structured result-set into fraud analysis Machine learning for key patternsBusiness benefit Avoid cost Improve margins Competitive pricing25 Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice.