Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Business Intelligence Industry Perspective Session I


Published on

Presentation by Kishaloya Roychowdhury and Koushik Roy

Published in: Education, Technology, Business
  • Be the first to comment

Business Intelligence Industry Perspective Session I

  1. 1. Business Intelligence ……industry perspective Kishaloya Roychowdhury Koushik Das
  2. 2. Background : - <ul><li>IT Enablement came into existence targeting improvement of enterprise operations through </li></ul><ul><ul><ul><li>Automation </li></ul></ul></ul><ul><ul><ul><li>Decreasing delays </li></ul></ul></ul><ul><ul><ul><li>Increasing accuracy & reducing ‘rework’ </li></ul></ul></ul><ul><ul><ul><li>Reducing cost </li></ul></ul></ul><ul><ul><ul><li>Providing more room to explore new ways of revenue </li></ul></ul></ul><ul><li>In older days </li></ul><ul><ul><ul><li>Business was less complex (geography bound, easier needs, limited & known customer base , less ‘ competition’ , less ‘ regulations ’, less diversity in ‘ product landscape’ etc.) </li></ul></ul></ul><ul><ul><ul><li>Information volume was less & could be manually managed </li></ul></ul></ul><ul><ul><ul><li>Simple ‘management’ like Cost & Profit was enough to run a business </li></ul></ul></ul><ul><ul><ul><li>It was difficult to envision the ‘ outcome ’ </li></ul></ul></ul><ul><li>IT systems were mostly operational systems </li></ul><ul><li>Management Information System used to replace manual management reporting </li></ul>
  3. 3. MIS Reporting - Overview <ul><li>A management information system ( MIS ) is a subset of the overall internal controls of a business covering the application of people, documents, technologies, and procedures to solve business problems such as costing a product, service or a business-wide strategy </li></ul><ul><li>Management information systems are distinct from regular information systems in that they are used to analyze other information systems applied in operational activities in the organization. </li></ul><ul><li>It has been described as, &quot;MIS 'lives' in the space that intersects technology and business. MIS combines technology with business to get people the information they need to do their jobs better/faster/smarter. </li></ul><ul><li>Old MIS Systems </li></ul><ul><li>Generally a few summary reports and a few detailed reports grouped for a business function and menu having multiple such groups </li></ul><ul><li>No well thought of framework for organizing, automating and analyzing business methodologies, metrics, processes and systems that drive business performance </li></ul><ul><li>Difficult to figure out ‘cause and effect’ relationships </li></ul><ul><li>Manual detection of problem points from a group of detailed reports </li></ul><ul><li>Decision making more dependent on intuition </li></ul><ul><li>New Generation MIS Systems (also termed as performance management systems) </li></ul><ul><li>Based on a sound framework for organizing, automating and analyzing business methodologies, processes and systems that drive business performance. </li></ul><ul><li>Business processes are aligned with Strategies and KPIs are aligned with business processes </li></ul><ul><li>Status indicators (KPI) set with defined target and/or tolerance ranges </li></ul><ul><li>KPIs are published into a dashboard / scorecard with the ability to drill down to detailed analysis or trend reports. </li></ul>
  4. 4. Importance of Reporting & Analytics <ul><li>Common needs of reporting & analytics from ages – </li></ul><ul><ul><ul><li>Understand the health of the Business at any organizational levels </li></ul></ul></ul><ul><ul><ul><li>Informed decision making at tactical & strategic level </li></ul></ul></ul><ul><ul><ul><li>Regulatory Compliance </li></ul></ul></ul><ul><li>Today’s need under the backdrop of ‘global competition’, ‘economic rollercoaster’ </li></ul><ul><ul><ul><li>Optimized but cost effective operations </li></ul></ul></ul><ul><ul><ul><li>Differentiation in the marketplace </li></ul></ul></ul><ul><ul><ul><li>Revenue protection and sustainable growth </li></ul></ul></ul>Early adopters ride the wave BI ANTICIPATE TRANSFORM AWARE
  5. 5. BI – Some real life industry needs <ul><li>Retail – </li></ul><ul><ul><li>Customer Intelligence </li></ul></ul><ul><ul><li>Product Pricing & Store Optimization </li></ul></ul><ul><ul><li>Right budgeting </li></ul></ul><ul><li>Finance – </li></ul><ul><ul><li>Right channel adoption </li></ul></ul><ul><ul><li>Intelligent customer service </li></ul></ul><ul><ul><li>Reduce financial risk </li></ul></ul><ul><li>Healthcare – </li></ul><ul><ul><li>Right care at the right time in the right setting </li></ul></ul><ul><ul><li>Disease management & Case management </li></ul></ul><ul><ul><li>Removal of behavioral barrier of doctors, patients </li></ul></ul><ul><li>Cross Industry – </li></ul><ul><ul><li>Cost reduction & safe revenue – (waste management & innovative practice adoption) </li></ul></ul><ul><ul><li>Regulatory Compliance </li></ul></ul><ul><ul><li>Performance Management </li></ul></ul><ul><ul><li>Risk & Fraud Management </li></ul></ul>
  6. 6. Challenges faced in the Industries <ul><li>It’s a sea of information </li></ul><ul><li>IT landscape (hardware, software, application) grew too large </li></ul><ul><li>No enterprise-wide standard (process, information, IT) </li></ul><ul><li>Information is not ‘trusted’ </li></ul><ul><li>Cross business function information gathering is dependent on huge manual effort & error prone </li></ul><ul><li>Some common issues faced by decision makers </li></ul><ul><ul><ul><li>Data is scattered everywhere across our organization. Where do I look ? </li></ul></ul></ul><ul><ul><ul><li>It takes forever to get the information I need to do my job </li></ul></ul></ul><ul><ul><ul><li>When I do get it, it’s wrong </li></ul></ul></ul><ul><ul><ul><li>We have mountains of data, but I can’t figure out what’s important </li></ul></ul></ul><ul><ul><ul><li>It takes so long to get the data that I don’t have any time left over to analyze it </li></ul></ul></ul><ul><ul><ul><li>I want it to be easy to see my data in every possible combination. Just let me point and click my way to an answer </li></ul></ul></ul><ul><ul><ul><li>I want a historical view of the business or make future projections </li></ul></ul></ul><ul><ul><ul><li>How can I plan based on the lessons learned and future projection </li></ul></ul></ul>
  7. 7. KPIs : selected measures of business performance Carefully selected set of measures derived from strategies, goals and objectives that represents a tool to communicating strategic direction to the organization for motivating change. These form the basis to plan, budget, structure the organization and to control results. Customer Measures % Sales of New Products Customers Acquired Customer Satisfaction Customer Retention Financial Measures Market Share % Revenue from New Products Transportation costs (costs/mile) R & D and Human Resource New Product Introduction Management Skills Employee Turnover Internal Process Measures Product Time to Market Unit Manufacturing Cost Days Supply to inventory % revenue from new products It is the ratio of money gained or lost (realized or unrealized) on the product relative to the amount of money invested on the same Customer retention rates Among the total customer, what are customers who are staying with the company after specific period of time. Its generally calculated yearly basis Customer satisfaction Customer satisfaction is a measure of how products and services supplied by a company meet or surpass customer expectation, for a particular service/product line. Generally estimated through survey and using a scoring model on a pre-defined scale.
  8. 8. Conceptual View of Enterprise Business Intelligence Executives Departments, Managers Touch Points- Customer, Vendor, Partner, Organization Units BI Real time BI, Embedded Analytics, Operational Dashboard Tactical BI Strategic BI Operational BI “ Capture” Distributed Operational Systems, External Data Providers, Unstructured Data Sources “ Collate” Centralized Information Warehouse “ Deliver” Distributed Analytic Applications Operate Plan Strategize Dashboards, Reporting, Analysis, Planning, Budgeting, Mining & Predictive Analysis Scorecards, Reporting, Planning, Budgeting, Performance Management
  10. 11. Conceptual DW Definition <ul><li>Data warehousing is a program dedicated to the delivery of ‘Enterprise wide view’ of information which advances decision making , improves business practices , and empowers workers . </li></ul><ul><li>The components, or layers, include the following: </li></ul><ul><ul><li>Business Architecture </li></ul></ul><ul><ul><li>Information Architecture </li></ul></ul><ul><ul><li>Applications Architecture </li></ul></ul><ul><ul><li>Data Architecture </li></ul></ul><ul><ul><li>Technology Infrastructure </li></ul></ul><ul><li>What a EDW is NOT </li></ul><ul><ul><li>A single integrated database or computer application </li></ul></ul><ul><ul><li>Not a duplication of every piece of data that exists in the Corporation </li></ul></ul><ul><ul><li>Up-to-the-minute reporting environment </li></ul></ul><ul><ul><li>A place to clean-up source system data accuracy issues </li></ul></ul><ul><ul><li>A means to perform the data conversion process for legacy system replacement projects </li></ul></ul>
  11. 12. EDW versus Data Mart EDW Data Mart <ul><li>Integrated (shared definitions) </li></ul><ul><li>Supports standard corporate definitions </li></ul><ul><li>Feeds Data Marts </li></ul><ul><li>Highest level of required granularity </li></ul><ul><li>Resides in a single integrated environment </li></ul><ul><li>Subject specific </li></ul><ul><li>Can be made of one or many datasets and/or data cubes </li></ul><ul><li>Accessed by the business users </li></ul><ul><li>Generally summarized </li></ul><ul><li>May reside on various computer platforms and environments </li></ul>
  12. 13. OLTP (online transaction processing) vs OLAP (online analytical processing) <ul><li>Organized around applications </li></ul><ul><li>Nonintegrated data </li></ul><ul><li>Different key structures </li></ul><ul><li>Different naming conventions </li></ul><ul><li>Different file formats </li></ul><ul><li>No time series analysis </li></ul><ul><li>Data relationships constantly change </li></ul><ul><li>Changes are instantaneous </li></ul><ul><li>Limited history, 60-90 days </li></ul><ul><li>Place an order for a product </li></ul><ul><li>Look up price for a product </li></ul><ul><li>Apply discount </li></ul><ul><li>Assign shipper </li></ul><ul><li>Trigger inventory pick-list </li></ul><ul><li>Verify shipment of product </li></ul><ul><li>Create invoice for the product </li></ul><ul><li>Apply credit to sales representative </li></ul><ul><li>Organized around subject areas </li></ul><ul><li>Integrated data </li></ul><ul><li>Standardized key structures </li></ul><ul><li>Standardized naming conventions </li></ul><ul><li>Standardized file formats </li></ul><ul><li>Time series analysis </li></ul><ul><li>Data is static over time </li></ul><ul><li>Series of data snapshots </li></ul><ul><li>Snapshots create historical database, often greater than two years </li></ul><ul><li>What type of customers are ordering this product? </li></ul><ul><li>Who are my top 10% accounts? By name, by revenue, by profitability, by region? </li></ul><ul><li>What have been the product purchase patterns over the past three years? </li></ul><ul><li>How are these different by customer segments? By sales rep? By store? </li></ul><ul><li>Which shippers have the best on time delivery records ? </li></ul><ul><li>How does this vary by shipment size? By season of year? </li></ul>Data organization & integration Time Handling Usage Examples Essential for running the company Essential for watching the company OLTP OLAP
  13. 14. Information Transformation Information Knowledge Intelligence Operational System Data Warehouse Data Marts BI solutions
  14. 15. Data Sources <ul><li>ERP or Custom implementations supporting operational need: </li></ul><ul><ul><li>HCM </li></ul></ul><ul><ul><li>CRM </li></ul></ul><ul><ul><li>Sales & Marketing </li></ul></ul><ul><ul><li>Order, Inventory </li></ul></ul><ul><ul><li>Procurement etc </li></ul></ul><ul><li>Manual systems mainly either in skill areas or around niche business functions: </li></ul><ul><ul><li>Planning & Budgeting </li></ul></ul><ul><ul><li>Customer profiling </li></ul></ul><ul><ul><li>Sales Rep Incentive Calculations etc </li></ul></ul><ul><li>Third Party Data: </li></ul><ul><ul><li>Credit rating </li></ul></ul><ul><ul><li>Competitor data </li></ul></ul><ul><ul><li>Prescription data </li></ul></ul><ul><ul><li>Address & demographic data </li></ul></ul><ul><ul><li>Market research data etc. </li></ul></ul>
  15. 16. ETL vs EAI Areas EAI ETL Definition Technology solution that enables systems to communicate Process designed by users to extract, transform, and load data from one or more sources to a target data repository Performance Optimization System is aimed at reducing the response time for a single user request or update System is aimed at reducing total time to create the unified historical record Integration Applications Data Focus Operational & Strategic Operational Business Case IT, e-business, Better Workflow, Data entry once Business Intelligence, Decision making, large volume, complex transformation, data quality Time Real Time Batch (moving to real time) Data Transactional-small Historical-enormous Metadata Limited--Message metadata Rich--dimensional metadata Transformations Format oriented --Code supported Analytic, Joins, Aggregations, function & formulae based Volume Single transactions Messages/second (KB) Days or weeks of data Records per min (GB) Targets OLTP API Code supported Relational Structures Native connectivity Codeless Extracts Data Using API’s Directly from database or using application adapters System Admin Involvement EAI requires no system administrator involvement. Once implemented, EAI is a technology solution that is transparent to end users. ETL requires extensive system administrator involvement ENTERPRISE BUS EXTRACT TRANSFORM LOAD Metadata ETL Transformation EDW
  16. 17. Operational Data Store <ul><li>An operational data store (or &quot; ODS &quot;) is a database designed to integrate data from multiple sources to make analysis and reporting easier. An ODS is usually designed to contain low level or atomic (indivisible) data (such as transactions and prices) with limited history that is captured &quot;real time&quot; or &quot;near real time“. </li></ul><ul><li>According to Bill Inmon , the originator of the concept, an ODS is &quot;a subject-oriented, integrated, volatile, current-valued, detailed-only collection of data in support of an organization's need for up-to-the-second, operational, integrated, collective information.&quot; </li></ul><ul><li>In practice ODS tend to be more reflective of source structures in order to speed implementations and provide a truer representation of production data. </li></ul><ul><li>An &quot; ODS &quot; is not a replacement or substitute for an enterprise data warehouse but in turn could become a source. </li></ul>Operational Data Store Data Warehouse Characteristics Data Focused Integration from Transaction Processing Systems, A better integrated picture of source systems Subject Oriented, Integrated, Non-Volatile, Time Variant Age Of The Data Current, Near Term (Today, LastWeek’s) Historic (Last Month, Qtrly, Five Years) Primary Use Day-To-Day Decisions, Tactical Reporting, Current Operational Results Long-Term Decisions, Strategic Reporting, Trend Detection Frequency Of Load Real Time, Near Real Time, Twice Daily , Daily, Weekly Daily, Weekly, Monthly, Quarterly, Bi-Yearly, Yearly
  17. 18. Staging Area <ul><li>Definition: Staging Area is a temporary location where data from source systems is copied and processed before loading into the target system, most often a data warehouse. </li></ul><ul><li>Minimizing processing on source systems </li></ul><ul><ul><li>Extract only once </li></ul></ul><ul><ul><li>Proper timing of different extracts within source system schedules </li></ul></ul><ul><ul><li>Both table-centric and document-centric extraction can be applied as necessary </li></ul></ul><ul><li>Source data within own control </li></ul><ul><ul><li>Incremental </li></ul></ul><ul><ul><li>Delta identification (Inserts, Updates, and Deletes) </li></ul></ul><ul><ul><ul><li>Reduce record set to be processed </li></ul></ul></ul><ul><ul><ul><ul><li>From source systems </li></ul></ul></ul></ul><ul><ul><ul><ul><li>For downstream processes </li></ul></ul></ul></ul><ul><ul><ul><li>True delta: only those records that have truly changed </li></ul></ul></ul><ul><ul><ul><ul><li>CRC (Cyclic Redundancy Checksum) </li></ul></ul></ul></ul><ul><ul><ul><ul><li>Column by column comparison </li></ul></ul></ul></ul><ul><ul><ul><li>Challenges in true delta identification e.g. </li></ul></ul></ul><ul><ul><ul><ul><li>NULL comparisons (Null does not equal Null) </li></ul></ul></ul></ul><ul><ul><ul><ul><li>When only the column used to identify a source modification changes </li></ul></ul></ul></ul><ul><ul><ul><ul><li>Source system challenges </li></ul></ul></ul></ul><ul><ul><li>Freedom of storage format and abstraction </li></ul></ul><ul><ul><ul><li>Data format consistency, e.g. </li></ul></ul></ul><ul><ul><ul><ul><li>CCYYMMDD format for dates, Trim trailing spaces, NULL replacement, Data type conversions </li></ul></ul></ul></ul><ul><ul><ul><ul><li>Demoralization, Document-centric, Summarization </li></ul></ul></ul></ul><ul><ul><li>Audit trail </li></ul></ul><ul><ul><li>Data Quality </li></ul></ul>
  18. 19. Real time data needs <ul><li>The barrier between transactional systems — which run the business day to day — and decision support — which traditionally have engaged business intelligence issues around product, customer, and market trends — is fading away under the pressure of new and ever more demanding business scenarios in customer service, product distribution, and market dynamics </li></ul><ul><ul><ul><li>A call center agent who has a customer on the phone at risk of going to the competition has 15 seconds to turn the situation around. </li></ul></ul></ul><ul><ul><ul><li>Analytics are used to optimize operations . For companies like FedEx, package dynamics are not just transactional; they are critical path — literally in a strategic and tactical sense — requiring the redeployment of resources such as aircraft to optimize operations. </li></ul></ul></ul><ul><ul><ul><li>Supplier scorecards — on-time deliveries, returns, defects, incomplete orders — reduce revenue losses from out-of-stock items and reduce markdown losses from overstocks. </li></ul></ul></ul><ul><ul><ul><li>For those enterprises that have physical inventory, reducing inventory through a demand planning or forecasting data warehouse results in significant cost reductions. </li></ul></ul></ul><ul><li>Data updates can only be as fast as the business processes that produce data. </li></ul><ul><li>Data consumption can only be as fast as the warehouse. </li></ul>
  19. 20. Right Latency is the right thing to implement Type Definition How it works Example Simulated An end user at a work station executing self-service query and reporting or what-if analysis. Updates and roll-up calculations are performed in batch, delivered in interactive “think time.” The results have been pre-computed and stored in the data warehouse for latter delivery as if the calculation were done in real time, but it is not. Customer recommendation Right time A catch-all phrase meaning near-real time — tied to a specific technology such as change data capture to a database log. Allows for a variety for response times, none committing to synchronous processing — allows for distribution by an ETL tool or message broker Web log analysis Real time The answer is absolutely the most up-to-date information physically possible in terms of both update and access. Resources such as databases, networks, and CPUs are locked synchronously until a commit point is reached, at which time other concurrent processing may proceed. Fraud detection On time Data is updated and delivered according to policies, service-level agreement, or consensus. Business groups tell IT how often they need to update and access data, and IT delivers data on that schedule. Inventory