Agile Business Intelligence
- 2. …“An approach that combines processes, methodologies, organizational structure, tools, and technologies that enable strategic, tactical, and operational decision-makers to be more flexible and more responsive to the fast pace of changes to business and regulatory requirements”. Forrester on Business Intelligence, 2011
- 3. • More than 49% of surveyed companies have concrete plans to expand their BI solutions by 2014 • Enterprise BI maturity levels are still below average (2.75 on a scale of 5, a modest 6% increase over 2009) • Lack of agility and flexibility leads to unsuccessful BI initiatives • Implementing BI requires using best practices and building upon lessons learned • BI technologies and processes have not kept pace with business realities • Centralization has not led to agile, streamlined BI implementations
- 4. • Agility is the key to efficient business intelligence processes • Untamed business processes create challenges for standard BI approaches: • Untamed processes are often invisible • Traditional SDLC doesn’t work well for automating untamed processes • Untamed processes often require new approaches and technologies
- 5. Business Intelligence Is Moving From Silos - To Centralization - To Agility 1990s 2000s 2010s Silos Centralization Agility Forrester, 2011
- 6. Business Process Example Why Agile process is needed Risk management Looking for trends of • Process involves complex transactions and exceptions salespeople violating discount • Correlation between discounts and individual contract exceptions policies may be unpredictable and on many different levels of a contract Sales Managing major accounts • Salespeople must compete against other companies for retailer shelf space and define the space their company has already secured • Salespeople may not know until they are in the client’s office what competing products the client is considering Sales and marketing Analysis of competitive new • Sales and marketing staff often needs to react to unexpected products, pricing, and competitive threats promotions • It’s difficult to predict all possible future competitive threats Paralegal Case support • Case workers often need to analyze multiple cases with disparate attributes for similarities and patterns • It’s difficult to predict all possible questions a lawyer might ask when analyzing a case Drug research Patient trials • Researching, investigating, and correlating drug treatment with patient outcomes requires a complex, multidimensional analysis with hundreds of potential cause/effect variables • It’s difficult to predict unexpected patient outcomes that will require a cause/effect analysis Forrester, 2011
- 7. Best practice Why it is important Recommendations Requirements for Some Business Processes Can Change Daily BI Business ownership and BI project ownership translates into • Demonstrate BI ROI to business leaders governance more successful BI environments • Run a BI maturity self-assessment and benchmark against peers and competitors Emphasize organizational Forcing decision-makers and • Foster a culture that makes change easier: set and cultural change knowledge workers to step outside of expectations up front, communicate often, collect management their comfort zones is a big change feedback, etc. • Build and use BI on BI • Make BI usage part of individual performance metrics, and even link it to compensation incentives Decouple data Data preparation requires more • Create separate, loosely integrated organizational preparation from data planning and control than data usage structures: put one in charge of data preparation, another usage processes in end- in charge of data usage to-end BI cycles • Emphasize IT ownership of data preparation processes Approach and treat front Front- and back-office BI applications • Create different sets of policies and guidelines for and back-office BI have different tolerance levels for approaching BI projects in the front and back offices Forrester, 2011 requirements and users risk, latency, planning, and data • Create special policies and guidelines for approaching BI differently accuracy projects that span untamed front- and back-office processes Establish hub-and-spoke Both extremes — organizational/data • Create a set of policies and guidelines that dictates which organizational models silos or a totally centralized BI data entities belong in the centralized (hub) area and which environment — have multiple negative belong in satellite organizations (spokes) implications • Base these policies on multiple parameters, such as how mission critical a data entity is or whether multiple units across the enterprise share its use
- 8. Best practice Why it is important Recommendations Use a combination of Neither approach is perfect: • Use a bottom-up approach for all data top-down and bottom-up • A bottom-up or horizontal approach preparation processes: sourcing, extraction, approaches to BI design requires building an enterprise data integration, cleansing, reconciliation, and modeling and applications warehouse and then applying it to • Use a top-down approach for all data usage reporting and analytical applications. processes: building reports and dashboards that link • A top-down or vertical approach clearly strategy to goals, goals to metrics, and metrics to links strategy, goals, and metrics to data data but often creates redundant efforts Use Agile development Traditional waterfall design and • Leverage Agile development best practices methodologies development methodologies are too slow • Support the Agile development methodology with an and too inflexible for BI Agile architecture and technologies Enable BI self-service for Even the best planning efforts can’t predict • Implement self-service BI tools and technologies business users future BI usage patterns • Ensure that at least 80% of all BI requirements can Forrester, 2011 be implemented by business users themselves
- 9. Automated Unified Automated information discovery Logically unified data sources Making BI contextual Data and content Full BI life-cycle automation Disk and streaming BI on BI Historical, current, and predictive Suggestive BI Complex data structures Automated decision management Metadata Agile BI Pervasive Limitless BI within processes BI within an information workplace (IW) Adaptive data models Self-service Unlimited dimensionality Mobile Exploration and analysis Offline/disconnected On-demand/SaaS
- 10. 1. Place BI agility and flexibility on an equal footing with risk management 2. Don’t fight people from the business who want to wrest control of BI away from you: rather, support them 3. Look for experience with next-gen and agile BI when selecting a systems integrator