Transcription
csBuIIAed Analyti2016e antropncIntellissegesinCapabilitiesR
EXECUTIVE SUMMARYAs organizations elevate their analyticscapabilities, one challengeof analytics maturity oftenoverlooked is the balance betweenstriving for more advanced capabilitiesand strengthening core businessintelligence competencies. Forexample, the decision to invest inreal-time supply chain optimizationmight come at the expense ofa supplier-facing website forperformance tracking. The centralquestion of how organizationsare approaching, managing, andsupporting the range of capabilitiesthat span Business Intelligence (BI) andAdvanced Analytics (AA) is the basis forthis report.IIAANALYTICS.COM 2016 IIA and SAS Institute Inc. All Rights Reserved.2
While the path from basic reporting to more advancedanalytics work is often considered as a shift from BIto AA, the reality is that advanced capabilities shouldaugment, not replace, less advanced functionality. Thisstudy looks at the interaction between BI and AA byprobing current-state attitudes and adoption of each,the many common and distinct barriers to success ineach, and the need to improve both. Our hypothesisis that viewing the two sets of capabilities through asingle lens will bring common pain points and clearrecommendations into better focus.Indeed, our hypothesis is strongly supported in thisreport. Using a survey of mid-market and enterprisescale companies, we’ve identified insights that, whencoupled with IIA’s experience in advising companies in allmatters of analytics, lead to specific recommendationsfor companies at different stages of maturity. The fourinsights and recommendations are as follows:O As expected, BI adoption is more prevalent than AA acrossboth mid-market and enterprise companies. This is somewhatexpected, and the implication is that AA requires dedicatedattention & focused investment. AA won’t happen organicallywithout its own initiatives.O For both BI and AA, the biggest measured gaps betweenimportance and performance are in data-related capabilities.The implication is that orgs must invest deeply in data,data platforms, and data-focused roles to unlock broadercapabilities across the full analytics continuum, not just tosupport AA.O For both BI and AA Weak Adopters, predictive analytics &leadership are the top identified gaps. Predictive analytics tellsus that companies recognize the ultimate goal and “North Star”of analytics, regardless of where they are currently. Our adviceis for companies to define strategic roadmaps with predictiveinsight, even if it’s in the distance. The gap in leadershipindicates companies need to take a hard look at who’s in place toshepherd the organization along the analytics path.“The silver lining in this report rests inthe challenges common to both BIand AA, such as data, leadership, andtalent: a rising tide driven by the rightinvestments and strategy should lift allboats across the analytics spectrum.“IIAANALYTICS.COMO The primary barriers to increasing analytics effectivenessinclude taking action, finding and retaining talent, improvinganalytics communication, and proving the value of analytics.This disjointed list of challenges resonates loudly with us atIIA, where we support companies that face an equally broadlist of barriers. Ultimately, companies must develop a trulymultifaceted strategy to drive sustained success and valuewith analytics. It’s not enough to invest just in data or solverecruiting challenges; all ingredients are important.Of course, these insights and words of advice come witha hefty price tag: competing on analytics requires thatorganizations take aim at, and invest in, the full analyticsspectrum in order to establish differentiating capabilities. 2016 IIA and SAS Institute Inc. All Rights Reserved.3
ABOUT THE RESEARCHThis research, commissioned bySAS and executed by theInternational Institutefor Analytics (IIA), sought tostudy the relationship betweenBusiness Intelligence (BI) andAdvanced Analytics (AA) inlarge organizations, and howend users can forge effectivepaths to increasing maturityand adoption.IIAANALYTICS.COM 2016 IIA and SAS Institute Inc. All Rights Reserved.4
Specifically, the research seeks to identify andunderstand current BI and AA adoption, the currentBI environment, capabilities and skillsets, and thepresence of self-service capabilities, all while exploringgaps and barriers to increasing organizationaleffectiveness of analytics capabilities.The followingis a breakdownof surveyrespondentsacrosscompany sizeand revenue:Results were drawn from 308 survey respondents,who were:O Employed full-time by a U.S. mid-market ( 50M to 1B revenue;n 105) or enterprise ( 1B revenue; n 203) company, across arange of industries,O Influencers in a BI/AA tool selection and adoption,or users of the BI/AA tool, andO End-users of the top 10 BI/AA tools like Excel, Oracle, IBM(Cognos, SPSS), BusinessObjects, SAS, Microsoft BI, Tableau,Qlik, Statistica, Tibco/Spotfire, MicroStrategyNUMBEROF RESPONDENTSCOMPANY SIZEMid-market500 to 4999 employees; with 50 million to less than 1 billion in revenue105Enterprise5000 employees; with 1 billion or more in revenue203Total308Respondents represent a range of industries and functions, including (but not limited to) financial services,banking, insurance, manufacturing, IT/technology, retail, and health care.IIAANALYTICS.COM 2016 IIA and SAS Institute Inc. All Rights Reserved.5
For the purposes of this survey, Advanced Analytics(AA) is defined as “the analysis of all kinds of datausing sophisticated quantitative methods (forexample, statistics, descriptive and predictive datamining, simulation and optimization of prescriptivesolutions) to produce insights that traditional BusinessIntelligence – such as query and reporting – are unlikelyto discover. Organizations commonly apply AdvancedAnalytics to data to find opportunities, mitigate risks,product or service innovation, acquire customers, andimprove operational effectiveness.” 1The term Business Intelligence (BI) includes “thereporting of historical and current business data to,for example, produce static reports, respond to ad hocrequests, provide for all online analytical processing,supply dashboards of key enterprise statistics, andserve other reporting needs.” 2THIS STUDY SEEKS TO PROVIDE ANSWERS TO THE FOLLOWING QUESTIONS:1Where areorganizations at intheir adoption of BIand AA, and how dothey execute the twodisciplines?2How do organizationsview their current BIand AA capabilities,and where are thelargest gaps betweenimportance andperformance?3What are the keythemes and barriersthat hamper adoptionof BI and AA?4How do organizationsplan to invest in BIand AA over the nextfew years, and whichemerging capabilitiesmight be implemented?1 Throughout the report, Advanced Analytics will henceforth be referred to as AA.2 Throughout the report, Business Intelligence will henceforth be referred to as BI.IIAANALYTICS.COM 2016 IIA and SAS Institute Inc. All Rights Reserved.6
1ADOPTIONAND EXECUTIONAs expected, Advanced Analytics lagsbehind Business Intelligence in termsof usage across an entireorganization. This lag is reflected inthe differences between several aspectsof BI vs. AA adoption, such as criticalityto business, recognition of benefits,and utilization in strategy. Inaddition, organizations strong intheir adoption of BI and AA are morelikely to have a central IT-led dataand analytics environment.IIAANALYTICS.COM 2016 IIA and SAS Institute Inc. All Rights Reserved.7
Advanced Analytics, however, tells a different story.There is not a significant gap in AA adoption betweenmid-market and enterprise organizations today (Figure1). Overall, 25% of enterprises use AA across the entireorganization in comparison to 23% of mid-market firms.Figure 1“The similarity in AA compared to BIreflects the challenges large enterpriseshave faced in building broad-basedanalytics usage, despite the establishedbase of BI capabilities in the organization.“Overall, enterprise organizations appear to be ahead ofmid-market firms with respect to BI adoption (Figure 1).Over half of enterprise organizations (52%) stated thatBI is used across the entire organization, compared to34% of mid-market organizations that said the same.However, mid-market firms are more likely to reportthat BI is used by many areas or is in the process ofexpanding to all areas of their organization (51% midmarket vs. 38% enterprise).Use of BI Within Organization38%Enterprise (BI)(n 203)10%51%Mid-Market (BI)(n 105)52%34%15%51%Enterprise (AA)(n 203)24%25%48%Mid-Market (AA)(n 105)29%Used by a fewindividuals/small number ofdepartments or areasIndicates significance at the 95% levelIIAANALYTICS.COM23%Used by many,but not all areas/in the process of expandingto all areas of organizationUsed across ourentire organizationPearson’s Correlation (r) Between BI and AA use across entireorganization 0.432, indicating strong positive correlation 2016 IIA and SAS Institute Inc. All Rights Reserved.8
Analyzing BI and AA adoption in conjunction with each other, we see that 41% of all companies that use BI acrosstheir entire organization also use AA across their entire organization (Figure 2). Consequently, companies that useBI across the entire organization are more likely to implement AA across the entire organization in comparison toother organizations, indicating the likelihood of common factors such as leadership, IT, and data investmentsdriving the adoption of both.Figure 2BI/AA Use Across Entire Organizationn 12Companies usingonly AA across theentire organizationn 77Companiesusing only BIacross theentireorganization8%51%n 62Companies using both AA and BI across the entire organization41%Throughout this report, we will explore the factors that drive BI and AA adoption, and the barriers to their adoption.We analyzed numerous survey questions by investigating the results in combination with BI and AA adoption,which were each defined as follows:Level of AdoptionLEVEL OF ADOPTIONStrong AdoptionBI (AA) is used across our entire organizationStrong AdoptionBI (AA)is usedbyacrossourbutentireorganizationBI (AA)is usedmany,notall areas, or is in the process ofIntermediate Adoptionexpanding to all areas of the organizationBI (AA) is used by many, but not all areas, or is in the process of expanding to allIntermediateAdoptionWeakAdoptionBI (AA)is ofusedby a few individuals or in a small number of areasareasthe organizationWeak AdoptionIIAANALYTICS.COMBI (AA) is used by a few individuals or in a small number of areas 2016 IIA and SAS Institute Inc. All Rights Reserved.9
First exploring the connection between adoption and criticality, firms strong in their adoption of BI are morelikely to indicate strong agreement with all statements related to BI use, growth and role in corporate strategy. Incomparison, firms with intermediate levels of BI adoption show a relatively lower level of agreement. Interestingly,companies with intermediate AA adoption are more likely to agree that AA is critical to their business, that its usein their organization will increase over the next year, and that employees outside of IT are able to make use of AAtools, as opposed to firms strong in their adoption of AA. As expected, firms weak in BI and AA adoption showedextremely low levels of agreement with respect to all statements relating to BI/AA use (Figure 3).Figure 3BI/AA Usage andImplementation (% Strongly Agree)Business Intelligence (BI)BI/Advanced Analytics is critical to ourbusinessThe use of BI/Advanced Analytics in myorganization will increase over the nextyearWe’ve seen significant benefits in our useof BI/Advanced AnalyticsWe use BI/Advanced Analytics to guidecorporate strategyIn my organization, employees outside ofIT are able to make use of BI/AdvancedAnalytics toolsAdvanced Analytics %9%61%48%35%47%4%5%65%37%30%5%Strong Adoption55%8%Intermediate AdoptionWeak AdoptionBusiness Intelligence: Strong Adoption, (n 139); Intermediate Adoption, (n 130); Weak Adoption, (n 34)Advanced Analytics: Strong Adoption, (n 74); Intermediate Adoption, (n 152); Weak Adoption, (n 77)Responses were on a 5 point scale, where 5 Strongly Agree, 4 Somewhat Agree, 3 Neither Agree nor Disagree, 2 Somewhat Disagreeand 1 Strongly DisagreeIIAANALYTICS.COM 2016 IIA and SAS Institute Inc. All Rights Reserved.10
This appears to indicate that the CIO or CTO of anorganization must have sufficient influence and/ora mandate to drive enterprise utilization, in orderfor many business functions within the company tomake it a priority.IIAANALYTICS.COM“In fact, in firms where AdvancedAnalytics has strong adoption, morethan three-quarters of the BI and AAdevelopment is led by IT, reflectingthe enabling role that a strong ITorganization can have on adoption. 2016 IIA and SAS Institute Inc. All Rights Reserved.“For both BI and AA, strong adopters are much morelikely to indicate that their centralized IT function ledthe implementation of data platforms and analyticstools, as opposed to a more decentralized approachwhere business users introduced them into theorganization and IT played more of a supporting role.11
2CAPABILITIES AND GAPSABoth enterprises and mid-market firmsplace a high importance on datarelated activities, such as preparingand cleaning data, but believe they arecurrently ineffective at it. As might beexpected, mid-market companies aremore likely to have these activitiesperformed by IT rather thanself-service.IIAANALYTICS.COM 2016 IIA and SAS Institute Inc. All Rights Reserved.12
For every BI and AA attribute tested, there existsan aggregate gap between stated importanceand performance, implying that the self-reportedeffectiveness of each capability was much lower thanthe stated importance among both mid-market andenterprise firms (Figure 4).Some of the largest gaps persisted among enterprises:in data-related activities such as filtering andtransforming data, preparing and cleaning data, andexploring data to identify causes. Among mid-marketfirms, visualizing data and building or updatingpredictive models using machine learning techniqueshad the greatest gaps.Figure 4BI/AA Attributes – Importance vs. PerformanceIMPORTANCEAttributesMid-market(n 105)BI ATTRIBUTESFiltering and transforming dataPreparing reports & dashboardsVisualizing data in an ad-hoc mannerPreparing and cleaning dataAverage Score of BI AttributesAA ATTRIBUTESGenerating large number of forecasts to improveplanning45%Analyzing unstructured data using text analyticsAverage Score of AA nterprise(n 203)28%50%32%37%47%36%40%54%41%38%Mid-market(n 105)49%50%Exploring data to identify root causes or trendsBuilding and updating predictive models usingmachine learning techniques(Top Box: % Very Effective)Enterprise(n 203)53%GAP SCOREPERFORMANCE(Top Box: % Very tems highlighted inindicate gaps where Top Box Importance scores were greater than Top Box Performance Scores by at least 20%Items highlighted inindicate gaps where Top Box Importance scores were greater than Top Box Performance Scores by at least 10% but no more than 20%IIAANALYTICS.COM 2016 IIA and SAS Institute Inc. All Rights Reserved.( Performance –Importance)Mid-market(n 105)Enterprise(n 24%-12%-8%-6%-7%-9%-14%13
A deeper look at the data provides more insight intowhat is critical from a performance perspective. Forexample, when looking exclusively at the proportionof those who rated a capability as ineffective while alsosimultaneously rating it to be very important, we seethat nearly one-quarter of large enterprises that placea high importance on preparing and cleaning databelieved their organizations were ineffective at it. Infact, filtering and transforming data was not as much ofa pain point – those that rated it as important appear tohave it figured out (Figure 5).Other opportunities include exploring data toidentify root causes or trends (among mid-marketfirms) and generating large numbers of forecaststo improve planning (among enterprise firms). Inline with the results observed earlier, this furtheraccentuates the fact that AA is behind BI in adoptionas well as performance.Furthermore, analyzing unstructured data using textanalytics emerged as another opportunity. While itscores lower in importance when compared to otherattributes, many of those organizations that placeda high value on it also believed they were currentlyineffective at it (Figure 5).Figure 5Opportunity Matrix – % of Those Reporting IneffectivenessAmong Those Who Rated Capability as Very ImportantINEFFECTIVENESS AMONG “VERY IMPORTANT”IMPORTANCE(% Not at all Effective/Somewhat Ineffective/Neither Effective norIneffective among those who mentioned “Very Important” for that item)(Top Box: % Very Important)AttributesMid-marketEnterprise53%49%BI ATTRIBUTESFiltering and transforming dataPreparing reports & dashboardsVisualizing data in an ad-hoc mannerAA ATTRIBUTESPreparing and cleaning dataGenerating large number of forecasts to improveplanningBuilding and updating predictive models usingmachine learning techniquesAnalyzing unstructured data using text analyticsItems circled withIIAANALYTICS.COM(n 56)8%(n 98)4%(n 52)6%(n 107)4%(n 47)15%(n 74)5%(n 43)13%(n 45)16%(n 79)17%(n 54)13%(n 111)28%13%(n 40)13%(n 55)29%16%(n 38)16%(n 58)54%37%47%41%40%43%Exploring data to identify root causes or n 94)indicate % where ineffectiveness was greater than 15% among everyone who rated that capability as “Very Important” 2016 IIA and SAS Institute Inc. All Rights Reserved.14
The following data-related activities are more likely to be primarily performed by IT in mid-market companies,compared to enterprises (Figure 6):O Cleaning and preparing data,O Visualizing data,O Exploring data to identify root causes or relationships,O Building predictive models using analytical techniques, andO Access and join data from multiple sources.In contrast, activities such as exploring data to identify root causes or relationships, and building predictive modelswere more likely to be self-service in enterprise firms than in mid-market firms.Figure 6Performance of BI and AA ActivitiesMid-market (n 105)Cleanse and prepare dataCreate dimensions, hierarchies, or aggregationsAccess and join data from multiple sourcesBuild predictive models using analytical techniques54%12%49%11%50%5%Exploring data to identify root causes or relationshipsCreate reports & dashboardsPrimarily by 48%11%Visualize dataIIAANALYTICS.COMEnterprise (n 203)23%20%38%34%Primarily Self-Service 2016 IIA and SAS Institute Inc. All Rights Reserved.15
BCompanies with weak AA adoption show the biggest gaps in execution of identified important capabilities;in contrast to BI, where lingering gaps persist across all levels of adoption. This highlights the intrinsicchallenges of strong BI execution.Further examination of perceived importance versus performance by level of adoption yields additional insightsfor both BI and AA. In the figure below, we see different types of gaps by activity according to the degree of BIadoption (Figure 7).Those with a strong level of BI adoption feel they are weakest when it comes to preparing and cleaning data, alongwith generating forecasts to improve planning; intermediate BI adopters are struggling more with filtering andtransforming data and exploring data to identify root causes/trends. Weak BI adopters are low performers acrossthe board, with their biggest challenge lying within the gap between importance and performance of preparingreports and dashboards.Figure 7BI/AA Attributes – Importance vs. Performance by BI AdoptionIMPORTANCEBusiness Intelligence (BI)BI ATTRIBUTESGAP SCORE(Top Box: % Very Effective)( Performance – Importance)Stro
(Cognos, SPSS), BusinessObjects, SAS , Microsoft BI, Tableau, Qlik, Statistica, Tibco/Spot!re, MicroStrategy The following is a breakdown of survey respondents across company size and revenue: Respondents represent a range of industries and fun
