Transcription

Proceedings of the International Conference on Industrial Engineering and Operations ManagementBangkok, Thailand, March 5-7, 2019Business Analytics Maturity Models: A Systematic Reviewof LiteratureKaveesha Ariyarathna and Suren PeterDepartment of Industrial ManagementUniversity of Kelaniya, Sri [email protected], [email protected] are increasingly using analytical tools to extract strategic information from data that have been collectedover many years, in order to derive a competitive advantage in the global market. The level of use of such tools vary,and different maturity models are used to represent different analytics implementations within organizations. Maturitymodels have been developed for both business analytics and business intelligence separately. Objective of this studyis to evaluate the maturity models developed recently. The business analytics maturity models consider both technicaland business aspects. A few models have considered both theoretical and practitioner knowledge. Nevertheless, a lackof consideration for theoretical knowledge in developing maturity models was noted. Most models consisted of stageswhere maturity is described under each level corresponding to the factors relevant to each stage. Lot of similar factorscan be seen among these models. These models act like a road map to achieve business analytics maturity withinenterprises. It was also found that the maturity models were developed in a developed country setting. A need for anintegrated maturity model with both technical and business aspects, incorporating the theoretical knowledge base andtaking factors affecting business analytics maturity in developing countries was established.KeywordsBusiness analytics, Business intelligence, Maturity models1. IntroductionBusiness Analytics refers to use of computing to get insight from data. It includes Business Intelligence and alsostatistical analysis. Collection, storage, analysis and interpretation of data is business analytics (Davenport and Harris2007). Organizations are using analytics in various types to make their business functions effective and optimal. Thesedata driven companies use predictive analytics, descriptive analytics, diagnostic analytics and prescriptive analyticsto solve day to day operations and run business in the long term. Analytics is primarily used in marketing relatedfunctions for understanding customer behavior and sales functions. Analytics is also used for financial relatedprocesses. Current development in technology especially with the advent of Industry 4.0 era, data and information hasgained attention from stakeholders in organizations. Analytics is not a novel concept but when organizations do notget the strategic information from available data using current technology that has been invested, it would be difficultto progress to the next stage of maturity. Business analytics also includes Business Intelligence which is howeversomewhat familiar area to organizations. Organizations who use analytics have been relatively more successful andgained competitive advantage, compared to others who do not use analytics. Use of analytics is not a one off exercise,but is a continuous process with varying factors coming into prominence depending at the level of maturity within theorganization. Both academics and practitioners have studied maturity in business analytics and developed BusinessAnalytics maturity models. The various models have wide similarities but some have distinctive features. However,most of them provide a roadmap to achieve maturity when using Business Analytics within organizations. It specifiesthe main factors that have to be addressed when moving up the maturity scale and how organizations should acquirethose capabilities. With the rapid development of information technology over the last three decades, businessanalytics has been intertwined with technology, where data and systems are involved. Some models have consideredboth technical and business views of Business Analytics while some have prioritized one aspect. The objective of thisstudy was to examine available Business Analytics maturity models including Business Intelligence maturity models.This paper provides a review of some of the more widely used business analytics maturity models and indicates theneed for an integrative model with which takes into account both business and technical aspects. IEOM Society International1762

Proceedings of the International Conference on Industrial Engineering and Operations ManagementBangkok, Thailand, March 5-7, 20192. MethodologyThe initial step of the analysis consisted of searching and reviewing of literature for Business Analytics and BusinessIntelligence maturity models. Fourteen related models or studies were found from Google Scholar, Emerald Insight,ResearchGate and IEEE Xplore Digital Library and an initial screening led to twelve of them being identified forfurther evaluation. Finally, a total of ten models were selected for evaluation. Although this search may not beexhaustive, it is believed that the papers selected and reviewed comprise a reasonable representative andcomprehensive body of the research work being accomplished in this area. Even there are industry specific BusinessIntelligence maturity models developments they were not considered for this study. The literature review wasconducted with the aim of identifying and revealing research gaps in the Business Analytics Maturity area. The reviewfocused on research institutes’ papers and conference papers within the period of 2007- 2018. Summary of themethodology is shown in Figure 1.1234 Web search (16) Initial screening (12) Review articles and assess for eligibility(10) Focus on 10 studiesFigure 1. Selection process of the study3. ResultsBusiness Analytics maturity models were studied mainly under two categories as Business Analytics and BusinessIntelligence maturity models.3.1 Business analytics maturity modelsBusiness Analytics maturity model developed by Gartner group is a more organization oriented maturity model. Theprogram management, technology and complexity of skills are associated with Business Analytics. Since all therequirements cannot be fulfilled at once, organizations need a framework to align Business Analytics through severalmethods. The main dimensions in the model are performance, people, processes, platform, program management,metadata and services. In order to improve organizational performance, it is imperative that enterprises measure theirperformances. Once organizational performance is measured, strategies should be changed or adjusted to minimizeperformance gaps. Then aligning analytics with these goals and objectives would be much easier. The model impliesthe significance of people in the organization where they act as producers, consumers and enablers of analytics. Inorder to increase efficiency and effectiveness in using analytics, the roles and abilities of these people must beunderstood and taken into account. Processes of organizations are categorized into three areas like decisions, analyticsand information governance. To get mature in analytics organizations should acquire capabilities relevant to eachthree processes. Decision capabilities imply collaboration, decision automation and various applications of analyticswithin the organization while using prescriptive, diagnostic, predictive and descriptive analytical capabilities. Datagovernance and integration with other systems like Enterprise Resource Planning systems, Customer RelationshipManagement systems and Supply Chain Management Systems are highlighted as information and governancecapabilities. Those systems assist enterprises to manage daily operations therefore integrity between them withanalytics cannot be neglected. Except that resource allocation on analytics, collaboration, and data sharing is alsoimportant in getting mature in analytics. Chandler et al. (2011) IEOM Society International1763

Proceedings of the International Conference on Industrial Engineering and Operations ManagementBangkok, Thailand, March 5-7, 2019TDWI’s analytics maturity model was developed by Halper and Stodder (2014) as an assessment tool to assessenterprises’ analytics capabilities. The model is described under five main dimensions such as organization,infrastructure, data management, analytics and governance. This model is more of a benchmark tool in businessanalytics and acts as a road map to gain maturity in analytics. Model is defined under five stages where organizations’analytics capabilities are shown in each dimension. There is a chasm between fourth and fifth stages. Strategy,leadership, skills, policies, data access, technologies and security and privacy are few of factors considered underabove five dimensions. Stages define how organizations move from one stage to next stage getting more capabilitiesin analytics. This model is a combination of both technical and business aspects in Business Analytics maturity.Davenport et al. (2010) developed the analytical maturity model DELTA. The model is named as DELTA because ithas five dimensions such as Data, Enterprise, Leadership, Targets and Analysts. It also describes the five stages ofanalytics maturity. These stages show how to initiate analytics in organizations and gradually move towards maturitygetting competitive advantage in the market. Data quality, integration, resource management, leadership support andorganization’s strategies are noted as the main factors at the different stages. This is more of a business view ofBusiness Analytics where technical aspect is considered a minor detail.Cosic et al. (2012) developed the theoretical based Business Analytics capability maturity model. This is developedbased on available maturity models in analytics and related functions. Model has four main capability areas and fourlow level capabilities under each main area. Business Analytics capability maturity in main areas leads enterprises tovalue and sustain competitive advantage. There is also a five level scale in maturity from non-existent to optimize.Main capability areas are governance, culture, technology and people. Under these strategic alignments, changemanagement, leadership, agility, system integration, data management and skills and knowledge are mentioned as lowlevel capabilities.3.2 Business intelligence maturity modelsHewlett-Packard (2007) has developed the Business Intelligence maturity model defining success of analytics as afunction of business enablement, information management and strategy and program management. In these threedimensions, advancement of analytics used in organizations for business needs, advancement in information solutionsand management skills needed are described. This model contains five levels where each gains maturity in thosedimensions. In these stages business enablement and strategy and program management dimensions’ maturity areexplained as a continuous process from running business to achieve excellence at the end through improvement andempowerment. Performance management, integration, strategic agility and shared resources are main factors in thismodel too. Even though Business Intelligence is more technical oriented this model can be considered as a morebusiness oriented maturity model.TDWI (2009) has developed a separate Business Intelligence maturity model where the Business Intelligenceimplementation in organizations evolves cost and value, to gain market share. It is developed in six stages in a curve,with a gulf and chasm. Dimensions in this model are type of system where the purpose of Business Intelligence isdescribed, analytical tools where Business Intelligence techniques organization is using is described and thearchitecture where data and information architecture is described. This model also indicates the maturity as frommoving driving the business to driving the market.Gartner has also developed a separate Business Intelligence maturity model where maturity of Business Intelligenceis expressed under three key areas people, processes, matrices and technology. This has five levels which is differentfrom Gartner’s Business Analytics maturity model. Moving from one maturity level to another requires changes in allcharacteristics that are, business model, management’s vision and data management. According to this modelachieving Business Intelligence maturity is not difficult. Wilen (2018)Chuah and Wong (2012) developed an Enterprise Business Intelligence Maturity Models to fill the gap betweenacademia and industry in Business Intelligence development. The integrated model developed following theCapability Maturity Model and consists of two representations; staged and continuous representation. There are fivestages as initial, managed, defined, quantitatively managed and optimizing. Continuous representation consists ofthirteen dimensions including: change management, organization culture, strategic management, people, performancemanagement, information quality and knowledge management. It highlights the importance of change management,organization culture, people, skills and data quality for Business Intelligence success. IEOM Society International1764

Proceedings of the International Conference on Industrial Engineering and Operations ManagementBangkok, Thailand, March 5-7, 2019Sacu and Spruit (2010) developed the Business Intelligence Development Model developed which contains six stages:predefined reporting, data marts, enterprise-wide data warehouse, predictive analytics, operational BI, businessperformance management with several characteristics categories like temporal, data, decision insight and outputinsight.Lukman et al. (2011) studied Business Intelligence maturity in Slovenian context considering BI in three differentsegmentations as technological, business and information quality viewpoints. According to the model, immatureorganizations should follow two paths to gain business intelligence maturity. Mainly organizations should focus onadvanced analytical technologies, data integration, fast access to information and fact based decision making. Modelis not exactly represented in stages but it indicates level development.4. AnalysisMaturity models studied under both Business Analytics and Business Intelligence mostly have similarities withdifferences in few areas. Three main factors should be taken into consideration. First, the perspective of the model.There are two insights, previous research has considered while developing those maturity models: technical insightand business insight. Some models have focused on both aspects equally while some have prioritized one aspect.Davenport et al. (2010) model, Hewlett-Packard’s model and (Chuah and Wong, 2012) model are more businessoriented models while (Sacu and Spruit, 2010) is more technical oriented. However, most of maturity models haveconsidered both aspects in maturity development since Business Analytics is consisted of both business and technicalperspectives. Second, the basis of developed model. The model can be developed based on theoretical background aswell as practical background. Studied maturity models have both approaches. Maturity model developed by (Sacuand Spruit, 2010) is based on theoretical knowledge while (Cosic et al., 2012) and (Chuah and Wong, 2012) are basedon both theoretical and empirical knowledge. Except that models developed by Halper and Stodder (2014) by TDWI,Chandler et al. (2011) by Gartner and Hewlett-Packard (2007) are more practitioner knowledge based. Thirdconsideration is the environment of the research conducted. The developed models constructs are based on developedcountries like the United States, European countries and in Australia. However, the contexts in developing countriesare very different from these countries. Developing countries may use the latest technologies similar to those in thedeveloped countries, but may be inhibited in realizing the full potential since the factors affecting Business Analyticsmaturity can be different. Table 1, shows a summary of models studied.Table 1. Summary of modelsModelChandler et al. (2011)Halper and Stodder (2014)Davenport et al. (2010)Cosic et al. (2012)Hewlett-Packard (2007)TDWI (2009)Wilen (2018)Chuah and Wong (2012)Sacu and Spruit (2010)Lukman et al. (2011)PerspectiveBusiness biasTechnical tagedmodel** IEOM Society International******************1765

Proceedings of the International Conference on Industrial Engineering and Operations ManagementBangkok, Thailand, March 5-7, 2019When we consider the factors used in the above models, it can be categorized into two as business related factors andtechnical related factors. As business related factors people, culture, performance management, business strategy,leadership and skills were noted. Data management, integration, quality, governance and technology were the mostcommon technical related factors. Another similarity among these models is many of them are stage developed modelswhere stages or levels are defined with different factors. Most of these models have five stages while two models(Sacu and Spruit, 2010)’s model and TDWI (2009) have six stages. Only Gartner’s Business Analytics modelChandler et al. (2011) is not stage based even though Gartner’s Business Intelligence maturity model Wilen (2018)has five stages.According to the summary above, it is noted that most models have taken a combined perspective including bothbusiness and technical aspects. Business Analytics is a combination of both of these perspectives. Therefore, focusingon both business and technical aspects are very important when assessing maturity. However, it was noted that manymodels lacked a theoretical base when deriving the models, even though some of the factors were similar. Therefore,consideration of practitioner knowledge is crucial to understanding the real word applications and issues. Further,these maturity models are generic and does not factor in differences in enterprise capabilities.5. ConclusionBusiness Analytics and Business Intelligence maturity models have been developed considering both theoretical andpractitioner knowledge. Many models have considered both business and technical aspects of Business Analyticswith different factors. Even though there are several models, there is still a deficiency of an integrated model withboth technical and business aspects that is derived from a theoretical base. None of these models is capable ofassessing all relevant factors to Business Analytics. Stage level models have defined the capabilities that should beacquired to gain maturity, though there is still no consensus in the method of assessing current maturity level oforganizations. Importantly all those models contain factors in a developed country setting which can be different to adeveloping country setting. Therefore, a need for an integrated model with both technical and business aspectsincorporating a theoretical base to assess Business Analytics maturity in a developing country setting, is identified.ReferencesChandler, N., Hostmann, B., Rayner, N., Herschel, G., Gartner’s business analytics framework, 2011, Gartner Inc,Available: usiness-intelligence/gartners business analytics 219420.pdf , June 29, 2018.Chuah, M. H., and Wong, K. L., An enterprise business An enterprise business intelligence maturity model ach-a-c; July 1, 2018.Cosic, R., Shanks, G., Maynard, S., Towards a business analytics capability maturity model, Australian Conferenceon Information Systems, Geelong, Australia, December 2012.Davenport, T. H. and Harris, J. G., Competing on analytics: the new science of winning, Harvard Business SchoolPress, 2007.Davenport, T., Harris, J., Morrison, B., Analytics at work: smarter decisions better results, Harvard Business SchoolPress, 2010.Halper, H., and Stodder, D., TDWI analytics maturity model guide, 2014, TDWI Research, b914607-084f-4869-ae64-e0b3f9e003de/TDWI Analytics-MaturityGuide 2014-2015.pdf, July 5, 2018.Hewlett-Packard, The HP business intelligence maturity model: describing the BI journey, 2007, ad.101com.com/pub/tdwi/Files/BI Maturity Model 4AA1 5467ENW.pdf, June 30, 2018.Lukman, T., Hackney, R., Popovic, A., Jaklic, J., Irani, Z., Business intelligence maturity: the economic transitionalcontext within Slovenia, Information Systems Management, vol 28, no. 3, pp. 211-222, 2011.Sacu, C., and Spruit, M., BIDM: the business intelligence development model, Proceedings of the 12th InternationalConference on Enterprise Information Systems, Madeira, Portugal, June, 2010. IEOM Society International1766

Proceedings of the International Conference on Industrial Engineering and Operations ManagementBangkok, Thailand, March 5-7, 2019TDWI, TDWI’sbusinessintelligencematuritymodel,2009, TDWI ntent/uploads/2013/05/TDWI BIMaturity0609 lettersize.pdf, July 5, 2018.Wilen, J., Business Intelligence, article 6: BI maturity model by Gartner, uha-wil%C3%A9n,October 10, 2018.BiographiesKaveesha Ariyarathna is an undergraduate at Department of Industrial Management, University of Kelaniya. She isreading for Bachelor of Science (Honours) degree in Management and Information Technology and major in BusinessSystems Engineering stream. Her interested research areas are in analytics, operations management and optimization.Suren Peter is a Senior Lecturer at the Department of Industrial Management, University of Kelaniya, Sri Lanka andhas served as the Head of Department previously. He received his doctorate and his Master of Philosophy degree fromMaastricht School of Management, The Netherlands. A Fulbright scholar, he graduated with a Master of Science inManagement from Georgia Institute of Technology in the USA. His first degree was in Industrial Management fromthe University of Kelaniya, Sri Lanka. He has served as a National Consultant on a number of projects for the UnitedNations (UNIDO and UNDP) and been involved in consulting with a number of international developmentorganizations, government and local private sector companies. He also serves on the Board of Directors of a numberof private companies engaged in finance and leasing, microfinance, real estate development and health and fitness.His areas of interest are in business analytics, capital market behavior, portfolio management and corporaterestructuring. IEOM Society International1767

Organizatio ns are using analytics in various types to make their business functions effective and optimal. These data driven companies use predictive analytics, descriptive analytics, diagnostic analytics and prescriptive analytics to solve day to day operations and run business in long term. the