Business Intelligence - Architecture & Execution Done Right
1 like815 views
The document discusses self-service business intelligence (BI), emphasizing its importance for empowering users to create reports independently, thereby improving decision-making processes. However, it highlights the necessity for a clear vision, executive support, and robust infrastructure to ensure successful implementation and maintenance. It also addresses challenges such as data accessibility, consistency, and user training, indicating that self-service BI is a dynamic process rather than a simple tool deployment.
Business Intelligence - Architecture & Execution Done Right
- 1. Self-Service Business Intelligence: Dispelling the Myth Presented to TDWI Northwest Chapter November 3, 2010 http://www.tdwi.org/
- 2. DATA DRIVEN DECISION MAKING: Holistically Designing & Executing Enterprise BI Decision Making at the Right Time and Right Level David Sogn Manager, Reporting & Analytics Disney Connected & Advanced Technologies
- 3. Self-Service Business Intelligence Anecdotal With Self-service BI , users create their own reports without having to rely on the IT department. Users get the exactly the reports they want, when they want them and the BI team no longer serves as an intermediary between user and the data TDWI Research Self-service business intelligence cannot simply be a portal or downloadable install from a serverencourages users to monitor and train each other, freeing IT staff to return to development Gartner Speedy answers, deeply dimensional data exploration, a single version of the truth, and a single view of all relevant business information Forrester Empowering the business user to access information they need, on-demand, without impact to IT - Infosys 3
- 4. Self-Service Business Intelligence Reality The ability to design, implement and execute an effective reporting & analysis rhythm, enabling decision making at the right time and level to improve short, mid and long term planning and business performance management Key Components: Ensure architecture stack highly scalable and flexible platform for collaborative, canned and ad-hoc BI over large data sets Segment users into distinct profiles while maintaining flexibility to evolve w/the business Prescriptive report & review process with clear purpose, in- and outputs of decision events Ensure predictable governance and support in place: Technical: Change request process for feedback and new feature roll-out. Operational: Leadership enforced follow thru and closing loop on decisions 4
- 5. Reporting Tools & Landscape: Ground Zero Problem Statement Information delivery and tools do not meet business requirements and needs Impact: Data I need for my role is not currently available I cannot make a decision Metric definitions, data and presentation are inconsistent and often inaccurate Impact: bad data + low confidence in data flow = decision paralysis (or worse, bad decisions) Initial development and ongoing production of reports is too time consuming Impact: more time spent building; less time spent analyzing/decision making Reconciliation between various publications is required Impact: data validation is time spent in the wrong area Tools and systems are instable and an insufficient support of business processes Impact: constant maintenance drives down productivity while various licenses and professional fees drive up costs 5
- 6. Enterprise DM&A Not A Hands Off Process Hypothetical Situation: Capture, Store, Map & Deliver Application Information Transaction Transformation Storage Reports (data source) Delivery 3 App AdHoc, Online Mapping Pivots Custom 1 DWH 3 6 App 5 4 2 Standard, Manual 5 Reports Validation 1 Scripts AccessDB Billing 2 2 Invoice Transaction (e.g. Business Apps and Transformation of Data Warehouses Excel Pivots and Excel report, DESCRIPTION order or journal online customer/Ad data, hierarchy (e.g. SQL, OLAP, MS Workbooks dashboard (web svc) entries) and raw Apps (e.g. SAP, mapping and Access, etc.) (reports), Data and PPT preso , data; manual input billing platform, Ad combing data feeds Extracts (e.g. MMX, hosted on MOSS or online delivery, Pricing, NNR, Omni, etc.) mash-up etc.) 1) Booking 2) Querying & 3) Adjustments & 4) Filling Gaps & 5) Dimensions & 6) Metric definitions Practices & Data Filtering Restatements Data Manipulation Hierarchy Mgmt and calculations Entry Methodologies -Process for -How are - Multiple instances -Definitions ISSUE TYPE - Entered correctly, -How is data pulled restatements? corrections made? of the same consistent across timely, consistent? from source? -How far back and - Gaps in data entry dimension exist? reports Measurement & -Bus. rules or filter? what implications on and measurement - What and who - Who approves Instrumentation -3rd Party treatment baseline or target? filled manually? owns master? changes - Consistent tags? 6
- 7. Reporting Tools & Landscape: Next Level Reporting Layer: Variety of data analysis and reporting tools Info Delivery: Single platform pulling from multiple sources Storage: Scalable, reliable storage of data with well managed security Transformation: Single place for data manipulation and hierarchy mappings Application: Adjustments and restatements done in in source application Transaction: consistent data entry and measurement practices 7
- 8. Ideal Technology Stack Business User Needs Technical Implications Easy, centralized access to Unified portal for all org Access all BI reports/tools relevant access to to users work reports, knowledge mgmt & support Reporting Personalized homepage based on users Analytical engines Agile access to most role/preferences relevant reporting, analysis Ability to dynamically slice and data needs for each property & dice data in multiple Property-specific aggregated data ways Frequent data uploads Aggregated data Cleaned-up, filtered User/role cluster centricity consistent data for most Single sources of truth Transactional data frequent business needs (taxonomy/hierarchy) Detailed and persistent data collection across properties Highly scalable log store* Property 1 Property 2 Property 3 Effective platform for Real time analytics short-cycle iteration on Real time Production multiple feature ideas feedback 8
- 9. Mapping Business Needs Using Innate Role Groups Major Business Unit LoBs Detailed, persistent data collection and storage Executive across properties is critical Platforms to overcoming siloed view Leadership 1. Smooth Deployment* 2. Coordination amongst roles and groups* Strategy Ecosystem General (field) Managers Role Groups Profile 1 Profile 2 Profile 3 Profile 4 Executive Leadership Corporate Leadership BU Lead Ops Lead Other (PMO, IT, Major Business Unit LoBs Channel/Site Mgr Dev/Planning Strategy/ Mktg Content Mgmt/Release General (field) Managers County Manager Biz Dev Manager Market Research Mgr Strategy Industry Research Competitive Research Domestic Strategy International Strategy Platforms Ad System Web Analytics Identity Mgmt /CRM Ops Infrastructure Ecosystem Biz Dev - Distribution Biz Dev - Syndication 9
- 10. Mapping Info. Needs to Specific Bus. Profiles Executives Managers Analysts Access to a dashboard for Ability to rapidly test new Ability to perform more monitoring key business ideas to succeed (or fail) complex analysis via enhanced metrics on a consistent and quickly query and drill-down timely basis capabilities Access to more insightful Access to more frequent Ability to create ad hoc data with linkages to root and insightful data allocating reports quickly and easily causes, allowing for faster and for better day to day decision more informed decision making making Ability to make most decision Improved ability to monitor Consistent view of data on data generated by current performance and available to do custom analysis predictably published standard initiative impact reports Ability to gain more insight Ability to answer most into consistent and well business questions rapidly by understood cross-property using flexible interfaces data and interactions 10
- 11. Wrap-up In summary: Self-service BI is not simply about providing customizable reports to employees and assuming that unfettered access to data will set them free. Quite the opposite, self-service BI is a living breathing process that requires: 1) a clear vision, 2) executive support 3) comprehensive roadmap/charter 4) exhaustive maintenance and support of the underlying infrastructure and data flows 5) customized training and socialization process for each specific user profiles. 11
Editor's Notes
- #8: * Canned 3rd party SaaS reports for Execs, Internal report builder for productManagers and unfettered access to aggregated data for Analysts
- #9: *All properties and markets; permanent append only
- #10: Smooth Deployment training for all levels (see Specific Business Profiles slide)Coordination amongst roles and groups- Intra-group: Roles w/in group use same data- Cross group: Fewer reports / data sources to understand use