The term, Big Data, was coined by Michael Cox and David Ellsworth in a 1997 article for a conference on visualization. The term was more about the size of the data sets that taxed the memory of computer systems. When data didn’t fit in memory, the solution was to get more resources. It is the first article in the ACM digital library to use the term “big data.”
Three Vs of Big data were first mentioned by Doug Laney in 2001 Volume – amount of data is massiveVelocity – speed at which data are being generated is very fastVariety– different types of data like structured and unstructuredVeracity – data must be accurate and truthfulVolume – humans create 2.5 quintillion bytes of data daily, 90% of today’s data has been generated in the past two years. created and machine createdVelocity – a study by 360i found that brands receive 350,000 Facebook likes per minute Velocity – 600,000 tweets per hour
The term,Big Data, is more of an amalgamation of different areas that help us try to get a handle on, insight from and use out of data. The focus needs to be on the data and less about the big. Data has always been big. I wrote my first book on a Mac Plus. Each chapter required its own floppy disk.
Kate Crawford, a Microsoft researcher and MIT professor, said that 2013 will be the year in which we reach the peak of the Big Data hype.According to Gartner’s hype cycle of emerging technologies,Big Data is headed toward it’s peak of inflated expectations and won’t reach the plateau of productivity for 2 to 5 years. The Plateau of Productivity represents the time when the technology finally delivers predictable value. The promise of Big Data, of course, is a treasure trove of high value across many industries – including healthcare. Everything from predictive and prescriptive analytics to population health, disease management, drug discovery and personalized medicine (delivered with much greater precision and higher efficacy) to name but a few.
Accenture surveyed 600 executives from the US and UK about their use of analytics. They found that the adoption of analytics is up and continues to grow, ROI remains elusive. Strategies usually are about making decisions. And when we make a decision, we typically eliminate an alternative course of action. Tactics are usually much more flexible. Strategies are about “what” we choose to do. Tactics are about “how” we choose to do it. It is often easier to change the “how” we do things than the “what”.Strategies are the investments of resources that build and grow an organization. Tactics are the day to day actions that get us to our goals.
When they linked their customer feedback to operational metrics, they got more value from the data as measured by executive satisfaction with the program and higher customer loyalty rankings within their industry.
Linkage analysis helps answer important questions that help senior management better manage its business. 1. What is the $ value of improving customer satisfaction/loyalty?2. Which operational metrics have the biggest impact on customer satisfaction/loyalty?3. Which employee/partner factors have the biggest impact on customer satisfaction/loyalty?The bottom line is that linkage analysis helps the company understand the causes and consequences of customer satisfaction and loyalty and thereby helping senior manager better manage its business.Understand the causes and consequences of customer satisfaction/loyalty
Avoid putting people on projects who are vested in the old way of doing things.
Kate Crawford from MITBe concerned about common method variance – things are correlated only because they are measured by the same thing
Customer Loyalty Measurement ApproachesObjectiveTime spent on web siteNumber products/servicespurchasedRenewed contractSubjective (self-reported)Likelihood to recommendLikelihood to continue purchasingLikelihood to renew contractHere is a figure that illustrates how different loyalty metrics fit into the larger customer loyalty measurementframework of loyalty types and measurement approaches. It is important to point out that the subjective measurement approach is not synonymous with emotional loyalty. Survey questions can be used to measure both emotional loyalty (e.g., overall satisfaction) as well as behavioral loyalty (e.g., likelihood to leave, likelihood to buy different products). In my prior research on measuring customer loyalty, I found that you can reliably and validly measure the different types of loyalty using survey questions.– I conducted several studies a few years ago, examining different types of customer loyalty questions. As the business models suggested, I found that loyalty questions generally fall into three types of loyalty behaviors. The table here includes these three types (Retention, Advocacy and Purchasing) and some survey questions for each type of loyalty.likelihood to switch providers (retention)likelihood to renew service contract (retention)likelihood to recommend (advocacy)overall satisfaction (advocacy)likelihood to purchase different solutions from <Company Name> (purchasing)likelihood to expand use of <Company Name’s> products throughout company (purchasing)These questions can be used as individual measures. If you are using multiple questions, you can calculate an average of the questions within a given type of loyalty. So, if you were using all six of these questions, you could calculate three scores, one for retention, one for advocacy and one for purchasing loyalty, each an average score of their corresponding questions.
Big Data - What it Really Means for VOC and Customer Experience Professionals
How may we firstname.lastname@example.orgSpring 2013Big Data: What it Really Means for VoC andCustomer Experience ProfessionalsBob E. Hayes, PhD
“Big Data” is Everywherehttp://www.evl.uic.edu/cavern/rg/20040525_renambot/Viz/parallel_volviz/paging_outofcore_viz97.pdf
Three Vs of Big DataINFOGRAPHIC from Domo June 2012VolumeVelocityVarietyhttp://blogs.gartner.com/doug-laney/files/2012/01/ad949-3D-Data-Management-Controlling-Data-Volume-Velocity-and-Variety.pdf
Big Interest in Big Data: Google Trends0204060801001202010-01-03 -2010-01-092010-06-20 -2010-06-262010-12-05 -2010-12-112011-05-22 -2011-05-282011-11-06 -2011-11-122012-04-22 -2012-04-282012-10-07 -2012-10-132013-03-24 -2013-03-30SearchVolumeIndexCustomer ExperienceBig DataScale is based on the average worldwide traffic of Customer Experience and Big Datafrom January 2010 to April 2013.
Big Data DefinitionAn amalgamation of differentareas* that help us get ahandle on, insight fromand use out of data* includes technology (Storage, Data Management, BI Reporting) and analytics
Big Data for Business – Getting Value5 High Value Use Cases*1. ExplorationFinding, visualizing and understanding alldata to improve business knowledge to makebetter decisions2. 360 degree viewof customerUnified view that incorporates both internaland external sources of customer data3. Security andIntelligenceDetect threat & fraud, Governance & RiskManagement, Monitor cyber-security4. OperationalAnalysisLeveraging machine data to improve results,Reduce resource costs5. Data WarehouseAugmentationAdding technology to data warehouse toincrease operational efficiencies and exploremore data.* Based on IBM’s review of over 100 real use cases: www.ibmbigdatahub.com/podcast/top-5-big-data-use-cases
Value from Analytics: MIT / IBM 2010 StudyTop-performingorganizationsuse analytics fivetimes more thanlower performershttp://sloanreview.mit.edu/the-magazine/2011-winter/52205/big-data-analytics-and-the-path-from-insights-to-value/
Value from Analytics: Accenture 2012 StudyCopyright 2013 TCELab1. Focus on Strategic Issues - only 39%said that the data they generate is"relevant to the business strategy"2. Measure Right Customer Metrics - only20% were very satisfied with the businessoutcomes of their existing analyticsprograms3. Integrate Business Metrics - Half of theexecutives indicated that data integrationremains a key challenge to them.http://www.accenture.com/us-en/Pages/insight-analytics-action.aspx
Data Integration is Key to Extracting Value96%72%51% 50%0%10%20%30%40%50%60%70%80%90%100%Percent of VOC executiveswho are satisfied withprogramCustomer loyalty percentilerank (within industry)PercentofVOCExecutives/CustomerLoyaltyPercentileRankOps Linkage AnalysisNo Ops Linkage Analysis
Data in Customer Experience Management1.Call handling time2.Number of calls untilresolution3.Response time1.Revenue2.Number of productspurchased3.Customer tenure4.Service contractrenewal5.Number of salestransactions6.Frequency ofpurchases1.Customer Loyalty2.Relationshipsatisfaction3.Transaction satisfaction1.Employee Loyalty2.Satisfaction withbusiness areasOperationalPartner Feedback1.Partner Loyalty2.Satisfaction withpartnering relationshipCustomerFeedbackEmployeeFeedbackFinancial
Integrate Data to Answer Different Questions• Linkage analysis answers the questions:– What is the $ value of improving customersatisfaction/loyalty?– Which operational metrics have the biggest impact oncustomer satisfaction/loyalty?– Which employee/partner factors have the biggest impact oncustomer satisfaction/loyalty?OperationalMetricsTransactionalSatisfactionRelationshipSatisfaction/LoyaltyFinancialBusinessMetricsConstituencySatisfaction/Loyalty
Integrating your Business DataCustomer Feedback Data SourcesRelationshipSurvey(satisfaction/loyalty tocompany)TransactionalSurvey(satisfaction with specifictransaction/interaction)Social Media/Communities(sentiment / shares / likes)BusinessDataSourcesFinancial(revenue, number ofsales)• Link data at customerlevel• Quality of therelationship (sat, loyalty)impacts financial metricsN/A• Link data at customer level• Quality of relationship(sentiment / likes / shares)impacts financial metricsOperational(call handling, responsetime)N/A• Link data at transactionlevel• Operational metrics impactquality of the transaction• Link data at transactionlevel• Operational metrics impactsentiment / likes/ sharesConstituency(employee / partnerfeedback)• Link data at constituencylevel• Constituency satisfactionimpacts customersatisfaction with overallrelationship• Link data at constituencylevel• Constituency satisfactionimpacts customersatisfaction with interaction• Link data at constituencylevel• Constituency satisfactionimpacts customersentiment / likes / shares
Selecting Your First Big Data Project• Identify/Discover all your data• Define your problem / Establish acompelling use case– Establish ROI• Select people before technology• Don’t introduce too many new skillsBased on: http://www.ibmbigdatahub.com/blog/selecting-your-first-big-data-project
Patient Experience Example – US Hospitals• Identify all your data (Medicare)– Patient Experience, Health Outcomes,Process Metrics, Financial• Define your Problem– What can hospitals do to improve patientexperience?– Does amount of hospital spend impact patientsatisfaction/loyalty?
Data Integration in US HealthcarePatientExperienceHealthOutcomesProcess(Operational)Financial• OverallSatisfaction• Likelihood torecommend• 8 dimensions• Nurse comm.• Doctor comm.• Room Quiet• MortalityRate• Re-admissionRate• SafetyMeasures• MedicareSpend• US Federal Government (Medicare) tracksseveral metrics across US Hospitals
Data Veracity – Accuracy and Truthfulness• Have ahypothesis(es)• Know whereyour datacome from• Consider theeffect size• Avoid cherrypicking results/ Be aware ofbiases
Problem of Common Method Variance• Correlations between variables are drivenby the method of measurement• Correlation between customer experienceand recommending behavior:– CX and Likelihood to recommend: r = .52– CX and Number of friends/colleagues: r = .28• Consider using objective loyalty metricswww.businessoverbroadway.com/is-the-importance-of-customer-experience-over-inflated
Customer Loyalty Measurement FrameworkLoyalty TypesEmotional BehavioralMeasurementApproachObjectiveADVOCACY• Number/Percent of newcustomersRETENTION• Churn rates• Service contract renewal ratesPURCHASING• Usage Metrics – Frequency ofuse/ visit, Page views• Sales Records - Number ofproducts purchasedSubjective(SurveyQuestions)ADVOCACY• Overall satisfaction• Likelihood to recommend• Likelihood to buy same product• Level of trust• Willing to forgive• Willing to considerRETENTION• Likelihood to renew service contract• Likelihood to leavePURCHASING• Likelihood to buy different/additional products• Likelihood to expand usage1 Using RAPID Loyalty Approach - Overall satisfaction rated on a scale from 0 (Extremely Dissatisfied) to 10 (Extremely Satisfied). Other questions arerated on a scale from 0 (Not at all likely) to 10 (Extremely likely). * Reverse coded so lower rates of these behaviors indicates higher levels of RetentionLoyalty. Copyright 2013 TCELab
Implications• Ask and answer bigger questions aboutyour customers– Explore all business data to understand theirimpact on customer experience / loyalty• Build your company around yourcustomers– Greenplum social network platform– Cross-functional teams working together tosolve customer problems• Use objective, “real,” loyalty metrics
email@example.com@bobehayesbusinessoverbroadway.com/blogHow may we firstname.lastname@example.orgSpring 2013Big Data: What it Really Means for VoC andCustomer Experience ProfessionalsBob E. Hayes, PhD