Quality information is key to making quick, rational business decisions BETTER DECISION MAKING WITH PROPER BUSINESS INTELLIGENCE | A.T. Kearney 1 A s companies focus on growth and business development, the availability of quality information is crucial to making quick, rational business decisions. New technologies, however, can cause information overload, leaving decision makers buried under a mass of irrelevant, inadequate, and inconsistent data. Some companies manage to provide their decision makers with precise, relevant data in an easy-to-grasp format. These companies have discovered the value of effective business intelligence capabilities. For many companies, new technologies are causing information overload, leaving decision makers overwhelmed with inadequate, incorrect, incon- sistent and misleading information. Indeed, the various acts of retrieving and processing this often useless information can tie up numerous resources. Yet there are companies that manage to provide their decision makers with processed and automatically consolidated raw data, presented in an understandable, easy-to-grasp format. These companies provide insightful information for quick, profound decision making. What do these companies have that the others dont? Business intelligence (BI) capabilities and processes. Business intelligence is a research eld that focuses on theoretical and practical aspects of achieving a solid information basis for decision making. This paper summarizes A.T. Kearneys experience in helping companies shape their data processes to obtain the right information for rational and quick decision making. What Is Business Intelligence? Business intelligence focuses on the particular eld of data processing and consolidation to retrieve information for decision making. The overarching objective is to providevia various solutions the right knowledge to the right people at the right time. Doing this requires the right mix of IT systems, architectures, data structures, data- collection processes, and responsibilities for provid- ing meaningful information. Business intelligence has a proven impact on key performance indica- tors (KPIs). For example: 60 percent of executive managers state that the use of a performance management tool has a positive impact on shareholder value Return on equity (ROE) is more than twice as high in companies that widely use perfor- mance management tools compared to those that do not in the same industry When assessing BI capabilities, there are four levels to consider: Reporting. Reporting is a core functionality of BI tools as the objective is to create recurring, standard, reports in an efcient and user-friendly manner. Reports are predefned and static by nature, generated either by request of an end-user or refreshed periodically through an automatic scheduler (uploaded on Intranet servers or shared BETTER DECISION MAKING WITH PROPER BUSINESS INTELLIGENCE | A.T. Kearney 3 BETTER DECISION MAKING WITH PROPER BUSINESS INTELLIGENCE | A.T. Kearney 2 Three Phases of Implementation A successful corporate BI implementation has three phases (see gure 1). In implementing BI sys- tems, it is important to begin with business requirements, because projects that are technically triggered usually fail. Phase 1: Dene the necessary KPIs on a management dashboard. A key activity in the rst phase is to dene future reports and KPIs, which often means eliminating some existing reports, as many are unnecessary and do not truly reect the companys objectives (see gure 2). 1
Next, is assessing data availability to calculate dened KPIs; a sound data basis is a key success factor for system implementation. External tools connecting to existing data warehouses and addi- tional calculations should be kept to a minimum. Also, the various layers of a BI system are evalu- ated to dene which areas should be addressed and identify specic parts that should be elimi- nated (see gure 3 on page 4). It is important to clarify at the beginning which areas shall be covered by the project: Is it a BI system for nance KPIs only or does it also include supply chain management? Will the proj- ect integrate other functional areas, departments or even specic business lines within operations? Often human resources and customer relationship management (CRM) are the next candidates to improve nance BI and integrate more of the drives and accessible to a predened group of cor- porate users). Key functionality is reduced to data consolidation and aggregation from various sources in a repetitive approach (automated, ide- ally) from trusted data sources. Dashboards. Dashboards contain high-level, aggregated strategic company data, inclusive com- parable presentations, and consolidated perfor- mance indicators. They include both static and interactive reports with data translated into graph- ics, charts, gauges and illustrations to simplify the communication of complex topics. Dashboards allow basic interactions (such as drill down, slice- and-dice operations to play with the data) and provide various levels of detail to achieve deeper insights. However, the explanatory power of dash- boards relies mostly on users interpretations. Analysis. At the analysis level, BI systems provide not only consolidated information that users can detail and lter, but also forecasts and trend analyses to develop new insights (based on the raw data). Analytics. At the top level of a BI system is automated intelligent data analyses based on sophisticated fuzzy logic and neuro-fuzzy sys- tems. Based on user-friendly but powerful func- tions, the BI system can retrieve meaningful insights while hiding the underlying complexity of data interpretation. What-if scenarios and simulation functionality provide advanced, tai- lored decision-making support. The fundamental basis of every BI system, however, is the data base on which it operates. Basic systems focus on corporate nancial data only, while advanced systems interconnect internal and external sources with qualitative and quantitative data. Advanced systems process the comprehensive data set using methods perfectly tailored to a rms needs and condense the ndings into meaningful knowledgehiding the complexity and selecting only the maximum amount of input necessary to help executives make the right judgments. The Advantages of Business Intelligence When analyzing a business intelligence solution, it is important to consider the business benets, including improved overall decision making and increased efciency for business reporting and analysis. To this rst point, BI offers four impor- tant prerequisites for proper decision making: Required information is available Data is consistent across organizational units Information can be easily analyzed using built- in analysis functionality Reports are presented in a user-friendly format A well-designed business intelligence solution ensures that information across the organization is available in a consistent, reliable manner. Figures can be aggregated and compared in different business units, assuring the validity of like-for- like data comparisons, and that all management reports provide operations leaders and top man- agement with the information they need to steer the business properly. Essentially, BI improves effciency on both the information technology (IT) side and the business side of the organization. On the IT side, workers are freed from the recurring task of creating and changing data reports as end-users are able to create and change their own reports. On the business side, less time is spent in data analysis and prepara- tion as management reports are created directly from the BI dashboards. Not only is the data in these reports more up-to-date and credible, but also they are easier to read and handle. And, impor- tantly, the information can be downloaded on smart devices, including the iPhone and iPad. The sidebar on page 4, What Makes a BI Leader?, offers a short list of success factors shared by top business intelligence organizations. Figure 1 Three phases of a business intelligence solution Source: A.T. Kearney analysis Phase 1 Define key performance indicators and level of business intelligence to be achieved Phase 2 Phase 3 Collect requirements, build prototype and conduct vendor proof of concept Implement solution and improve tool according to release plan Figure 2 The main tasks in Phase 1 Source: A.T. Kearney analysis Evaluate potential key performance indicators (KPIs) Analyze empiricial findings related to use of KPIs Evaluate simple profitability metrics Assess select KPIs for appropriateness Compare analysis of value-add KPIs Identify potential risks of making the wrong decision Perform qualitative evaluation Determine value correlations Analyze risks Make recommendations 1 Phase 1 should always be completed fully before proceeding with the second phase, as the required reporting and the defined KPIs set the framework for tool selection. BETTER DECISION MAKING WITH PROPER BUSINESS INTELLIGENCE | A.T. Kearney 5 BETTER DECISION MAKING WITH PROPER BUSINESS INTELLIGENCE | A.T. Kearney 4 business. The link to operations is more difcult as KPIs from manufacturing, production or logis- tics are fairly different from nance KPIs and are often difcult to connect to existing KPI trees. Even between similar business lines, KPIs are sometimes slightly different because they repre- sent different business models for instance, internal production for intermediate products or external production for nal products. If the long- term strategy is to analyze operational KPIs, this should be addressed at the outset. Once these questions are answered and all affected business areas are addressed, the second phase can begin. Phase 2: Create a design and navigation prototype and build a proof of concept. Four tasks are involved in preparation for system imple- mentation. The result is a proof of concept, designed before the comprehensive implementa- tion begins. At this point, software vendor selec- tion is independent of specic design requirements and often driven by strategic policies or IT land- scape requirements. Based on IT architecture rules, a short list can be devised up-front to select appropriate tools to t the companys require- ments, IT strategy and IT landscape (see gure 4). Phase 2 is not only about collecting require- ments regarding the functionality of a future tool (navigation and analysis deep dives, for example) but also about report design, dashboards and tool functionality. One point must be stated very clearly: A pure management cockpit or dash- board cannot replace the internal or external reporting. As a rst step, it can be seen as a second channel (always available), but with a different level of detail (management adequate). All man- agement cockpit tools offer a reporting function- ality used to print-out the dashboard content. Replacing the complete paper-based internal (or external) reporting requires signicant efforts as the detailed design of every single page is outlined up-front. Both the dashboard screen design and the report layouts, which cover all of the depicted content regarding historic numbers and compari- sons, among other things, nally get dened before the detailed concept is handed over to a system integrator for implementation. The real- ization of the screens and the corresponding KPI visualization are the main drivers of the system implementation and the testing. Therefore stabil- ity is required. Figure 3 Layers of a business intelligence system Sources: Forrester, A.T. Kearney analysis Delivery A p p s F o r m
f a c t o r P S O Reporting Performance management Support applications Analytics Discovery and integration Data Infrastructure Network Report mining DQ cleansing, profiling EAI / SOA EII ETL / CDC Discovery accelerators Adapters / tool kits Accelerators / query optimization BPM / BRE integration BAM / CEP Operational data stores (ODS), data warehouses (DW), data marts (DM) Integration third party applications Services registry and repository Streaming DBMS Search DBMS RDBMS In-memory DBMS Hierarchical / XML Columnar DBMS Multi-value RDBMS Multi-dimensional OLAP Usage analytics Statistical analysis Web analytics Data / text mining Guided decisions NLP Guided search Time series OLAP Operational DSS Predictive analytics Servers Storage MDM S I I n d u s t r y
v e r t i c a l
a p p l i c a t i o n s S t r a t e g y A p p l i a n c e M e t h o d o l o g y G o v e r n a n c e B I S a a S H o s t e d
B I
( A S P ) b - l a y e r s ) M S P
/
a p p s
o u t s o u r c i n g C e n t e r
o f
E x c e l l e n c e E n t e r p r i s e
a p p s :
E R P ,
C R M ,
S C M ,
E R M B P O eLearning ECM Metadata-integration, repositories Version control Reporting: ad-hoc, analytical, production Search Geospatial Advanced data visualization Alerts Dashboards Collaboration Life-cycle Mgt. Localization QA Strategy / objectives management Metrics / KPIs Planning Scorecards Interactive voice response, ATM, point-of-sale Portals Desktop gadgets Office suites Mobile Disconnected O r g What Makes a BI Leader? From our experience helping com- panies implement their business intelligence projects, we have found several common characteristics or success factors that differentiate the BI leaders from the followers. The BI leaders: Defne and optimize corporate report standards Focus on the business not the data Establish a balanced scorecard Defne clear data dimensions and structures Develop a master data manage- ment plan Establish corporate data gover- nance and clear ownership of master data Use report visualization in pilots Leverage web technologies Focus on cross-organizational efforts Distinguish between business- driven and IT-driven projects Defne the long-term BI scope, from the beginning Realize a mobile version of business intelligence Figure 4 The tasks in phase 2 Source: A.T. Kearney analysis Perform assessment to ensure the need for a cockpit Conduct a brief cost- benefit analysis of the area and other potentially affected areas such as procurement Draw up a short list of vendors, taking into account industry reports and other external information Assess the options from an internal viewpoint Complete a list of business requirements and build a first simple PowerPoint prototype to develop a look-and- feel of the final product Begin company-vendor discussions Help the vendor understand the client and ensure vendor receives timely feedback so to make quick improvements Assessment Vendor selection Requirements and prototypes Proof of concept BETTER DECISION MAKING WITH PROPER BUSINESS INTELLIGENCE | A.T. Kearney 7 BETTER DECISION MAKING WITH PROPER BUSINESS INTELLIGENCE | A.T. Kearney 6 The proof of concept increases the developers understanding of the companys requirements before starting the real systems implementation. In our experience, a quick prototype allows a company to check the design of all possible pre- selected tools and test the capability of the poten- tial solution provider. Work packages can also be tested to assess vendors innovation capabilities in solving design questions. Finally, the rms requirements can be tested and challenged. Phase 3: Implementation. Implementation is a typical IT project. However, due to high visibil- ity and management awareness, the implementa- tion should be fast and provide quick solutions. A release plan ensures rst results are delivered quickly, while the nal and comprehensive solu- tion is created in several steps, aligned with report- ing cycles and data availability. The rst release should provide all required dashboard functions (design, navigation and drill-down) and some of the reporting requirements, which are detailed in the follow-on implementation phases and can be delivered sequentially. Typically, rst releases lack the full data set. An essential activity in the BI project is to cleanse existing data and establish a process to record new data in a proper way, which often requires developing the entire data architecture, including dimensions, hierarchies and formulas. Close interaction between system provider and business departments (as the nal users and clients of the system) is crucial to the success of the project. The management cockpit or dashboard fullls the design and functional requirements because it provides clear visibility to management and the necessary ease of use. Although all required data or KPIs may not be included in the beginning, they are provided throughout the steps. Governance and the CIOs Role Within a BI project, the corporate IT and the CIO moderate between the different business departments that are involved, the internal IT group, and the software provider or implementer. A BI project is often driven by the nance depart- ment as the key user due to corporate reporting and corporate management requirements; nance denes the main KPIs on a corporate level and ensures standardization across all different busi- ness models within the rm. Operational KPIs are provided by the differ- ent operating units and standardized to ensure that consistent content is reported to top manage- ment. The CIO and corporate IT take over the role of supporting KPI denitions by providing information about data sources, data availability and data quality. And they start the analysis on how a business intelligence tool could be used to fulll high-level requirements (see sidebar: Business Intelligence: A Case Study). Part of the project involves dening future BI governance. This often means establishing a BI Business Intelligence: A Case Study The management team at a large German company decided to imple- ment a corporate-wide business intelligence (BI) process. The com- pany, involved in heterogeneous businesses, started with pilot units in its three business concerns, den- ing nancial key performance indi- cators (KPIs) to manage and direct the corporation. The management team knew that success would depend on cor- porate-wide denitions for its KPIs so they could be cascaded down to all business units. The project began with develop- ment of a basic management dash- board that provided a visual KPI tree to clarify the logical dependencies of the two main KPIs and then illus- trating all the subsequent KPIs that owed from themand how they navigated through the organization. Next, business-specic operational KPIs were added to the dashboard, a simulation tool was developed to provide units with a business- planning model to illustrate the dependencies between KPIs, and a BI tool-based reporting process was designed and implemented The IT work began during the nal denition of corporate-wide KPIs and reporting, providing infor- mation about the availability of exist- ing data within the data warehouse. This was followed by the vendor and tool-selection process. A proto- type was designed, and software pro- viders were asked to use the design and business units functional requirements to develop a rough proof of concept. Immediately after the proof-of- concept presentations, the vendor- selection process was nalized and implementation began. Within three months, the company had its rst release, covering main functional- ities, dashboard design and naviga- tion, and rst-reporting factors. The release plan took into account the main reporting cycles, and two sub- sequent releases were scheduled within the planned timeframe. Soon after introduction, the management team was delighted as user acceptance proved stronger than expected throughout the organiza- tion. When asked about the success factors, team members were quick to cite the top two: design and ease of navigation within the BI tool, and on-time delivery of the major KPIs. All in all, the company is now dedicated to using business intelli- gence to help steer it on a richer and more successful course. Through TOOL 1 WYSIWYG the dashboard can be handled Vendor 1offers functional building blocks for the storing of comments in the SAP BW Simulations can be saved; retraction through vendor 1s own functionalities into BW backend Look-and-feel remains the same Comments and simulation can be executed early Either DataMart on machine two or WebService on machine one BOE Infrastructure for LiveOffice and reporting tool (phase2) Possible according to the demonstration; no information given on the interface builder; use of all graphic types possible Possible according to the demonstration Comments can only be used in the machine-three environment Possible according to the demonstration; simulations can be saved and retracted into the SAP BW Look-and-feel remains the same Comments and simulation can be executed early According to vendor, it can communicate with SAP BW 3.5 but not demonstrated Basic infrastructure on machine one has to be built anew Figure 5 Detailed analysis of BI software vendors Source: A.T. Kearney analysis Full function Only static design possible; performance critical Use of standard BDS associated with performance issues Generally not simulation functionality Due to low performance on comments and the illustration of graphs, no usage possible in phases 1 to 2 Executed on machine one, no export of the cube data needed None Implementation design Implementation comments Implementation simulation Transition from prototype to release I to II Connectivity to data Hardware costs Category On-board tools Vendor 1 Vendor 2 Very limited function A.T. Kearney is a global management consulting rm that uses strategic insight, tailored solutions and a collaborative working style to help clients achieve sustainable results. Since 1926, we have been trusted advisors on CEO-agenda issues to the worlds leading corporations across all major industries. A.T. Kearneys ofces are located in major business centers in 37 countries. AMERICAS Atlanta | Boston | Chicago | Dallas | Detroit | Mexico City New York | San Francisco | So Paulo | Toronto | Washington, D.C. EUROPE Amsterdam | Berlin | Brussels | Bucharest | Copenhagen Dsseldorf | Frankfurt | Helsinki | Kiev | Lisbon | Ljubljana | London Madrid | Milan | Moscow | Munich | Oslo | Paris | Prague | Rome Stockholm | Stuttgart | Vienna | Warsaw | Zurich ASIA PACIFIC Bangkok | Beijing | Hong Kong | Jakarta | Kuala Lumpur Melbourne | Mumbai | New Delhi | Seoul | Shanghai | Singapore Sydney | Tokyo MIDDLE EAST Abu Dhabi | Dubai | Johannesburg | Manama | Riyadh & AFRICA For information on obtaining additional copies, permission to reprint or translate this work, and all other correspondence, please contact: A.T. Kearney, Inc. Marketing & Communications 222 West Adams Street Chicago, Illinois 60606 U.S.A. 1 312 648 0111 email: insight@atkearney.com www.atkearney.com 2011, A.T. Kearney, Inc. All rights reserved. A.T. Kearney Korea LLC is a separate and independent legal entity operating under the A.T. Kearney name in Korea. BETTER DECISION MAKING WITH PROPER BUSINESS INTELLIGENCE | A.T. Kearney 8 center of excellence led by the nance department that includes data governance with BI system maintenance authority. Corporate IT should be part of the change advisory board that assesses and evaluates change requests based on the capabilities of the system solution, while IT maintains tech- nology authority. Choosing Tools and an Implementation Partner BI tool selection is rarely a green eld approach. The corporate IT strategy, system landscape and the main requirements regarding the BI level to be achieved are all guidelines in the selection pro- cess. Market research and analysts estimations can be used as support documents, but in most cases these analyses are too general and cannot be adapted to meet company specics. Instead, external reports are typically used afterwards as support information to justify the short-list ven- dors or tools. The software vendor and implementer are the same in most cases, as only a few companies are able to provide a full business intelligence suite, connected or even integrated somehow to the main enterprise resource planning (ERP) systems, such as SAP or Oracle. A company-specic questionnaire and evalua- tion template allows the efcient gathering of key requirements. The main questions or key evalua- tion points should be shared in advance with the short-listed software vendors to give them an opportunity to prepare the tool presentation and to answer all questions completely. The clearer the requirements are documented and described, the faster the selection process can be executed (see gure 5 on page 7). A rough prototypesuch as a PowerPoint visual of the management cockpit to demonstrate key navigation functions and design-related requirements will help both sides understand the desired outcome and manage expectations. It is the basis for the later detailed design and concept, containing the description of each requirement. This and the more technical description of the existing landscape and necessary interfaces are the most relevant documents for the implementa- tion and build up to the kick-off for the system- implementation phase. BI: A Must for Business Success A proper BI solution is a must have in todays world. Companies in all industries are using BI systems for successful decision making. These companies beat their competition and identify new opportunities to optimize their businesses. They also reduce resources for manual effort and rededicate people to analyzing data and preparing decision memos. For companies that grasp the true potential in business intelligence, they should take action sooner rather than later time, as always, is of the essence. Authors Alexander Martin is a principal in the strategic information technology practice. Based in the Dsseldorf ofce, he can be reached at alexander.martin@atkearney.com. Robert Jekel is a consultant based in the Zurich ofce. He can be reached at robert.jekel@atkearney.com. Edgar Simons is a consultant in the Dsseldorf ofce. He can be reached at edgar.simons@atkearney.com. ATK.0211.172