Oracle Retail Data Model (ORDM) Overview May, 2014
Content Retail Business Intelligence Key Trends Retail Industry Findings Foundation for Business Information Flows Retail is being Redefined Challengers for Getting Actionable Answers Oracle Retail Data Model (ORDM) Overview Typical Issues Addressed by ORDM ORDM Subject Areas Oracle Retail Business Intelligence Space Complete Oracle Data Warehouse / BI Solution ORDM Technical Architecture ORDM Components Relationship Sample Analytical Reports Sample Data Mining Packages and Data Model Type 1 Q & A
Retail Business Intelligence Key Trends Based on Industry Research, Retail will be an area of stronger than average growth compared to other industries Retail is the second largest growing segment for Analytics and Information Management 2
Retail Industry Findings Within the trends of today s Retail Market, there are opportunities for growth in Retail Business Intelligence Trend Details BI Opportunity Customer Interaction Growth in the number of vehicles to interact with customers, including in-store kiosk displays, e-commerce, call centers, catalogs sales, and mobile devices Intelligence to monitor customer interactions with a retailer through promotion tracking and segmentation Customer Social Networking Brand loyalty Widespread adoption of social networking to allow customers to provide product reviews and feedback Need to increase brand loyalty and customer retention at a time of increased competition Analytical insight into highly regarded products and services from customers Identify specific customer segments and tailor marketing and brand strategy to serve those customers Customer Behavior Retail IT Budgets Capturing relevant, timely, and granular data on customer buying behavior and to provide one view of the customer Tight margins may inhibit investment in new technologies and processes. Identify customer activity across multiple channels and brands and create a total, 360 view of each customer, identify cross-selling opportunities Develop methodology accelerators to reduce implementation costs and total cost of ownership.. 3
Foundation for Business Information Flows Manufacturing/Sourcing Sales Forecasting Inventory Tracking 4
Retail is being Redefined Business Areas Loyalty Marketing Category Item Price Multi-Channel Sales Prospects & Customer Store Labor & Operations Forecasting & Scoring Insight Requirements What are the characteristics of my most loyal customers? Least loyal? How do customers feel about our company and products? Which items drive sales? Which items are frequently purchased together? If I discount an item, will impact will it have on sales and revenue? How do my internet sales compare to brick and mortar in terms of revenue and cost? Which prospects should I target to convert into loyal customers? What products or offers would be most effective? Which combination of employees maximized store performance? Will my inventory levels meet sales forecast? When will we run out of stock? 5
Challengers for getting Actionable Answers Data Latency From event to action, too many hoops Moving from batch oriented to event oriented Shared Semantic Understanding Common vs. Canonical Foundation Operational vs. Analytical Volume and Velocity Acquire and Filter Organize and Analyze Retention 6
Oracle Retail Data Model (ORDM) Overview - Introduction - 7
Oracle Retail Data Model (ORDM) Overview It speeds the development of a data warehouse solution by providing a foundation data warehouse and analytic infrastructure for the reporting needs of a retail operation The Oracle Retail Data Model (ORDM) is a start-up kit for implementing a retail data warehouse solution Based on ARTS 6.0 Extraction of detailed and summary data Summaries, trends, and forecasts Knowledge discovery of hidden patterns and insights Information Analysis Insight and Prediction Who purchased cigarettes in the last 3 months? What is the average income of cigarettes buyers by region, by year? Who will buy cigarettes and beer in the next 6 months and why? 8
Typical Issues Addressed by ORDM (1) How will a business benefit from using ORDM? ORDM enables business users to turn retail business data into information based on which users make decisions. ORDM uses pre-built mining models to detect hidden information in data repositories, which helps businesses in: Determining how are my product and Point-Of-Sale performing? Determining what is my gross margin return on space? Determining how is the business doing compared to last year? Compared to plan? Determining what are my potential out-of-stock situations? Determining if the product assortment is optimal for all my regions Retaining customers and avoiding churn Profiling customers to understand behavior Finding rare events Targeting customers with the right offer and thus reducing customer acquisition costs 9 Maintaining and improving profit margins
Typical Issues Addressed by ORDM (2) How will a business benefit from using ORDM data mining functionality? ORDM data mining packages can be used for: Market basket analysis Frequent shopper analysis Identify and predict best customers and undesirable customers Calculate the probability of a customer buying Product X Detect churn or the probability of a customer switching from brand A to brand B Determine best location traits for stores Shrinkage analysis Associate items in a market basket Sales performance prediction Employee performance 10
DB RAC Partition OLAP Mining Spatial Compression ORDM Subject Areas Oracle Retail Data Model Includes Data Model with 650+ Tables and 10500+ Data Model with 650+ Tables and 10500+ attributes Industry specific 1200+ Measures & KPI Pre-built 15+ OLAP cubes Pre-built 12+ Mining Models Complete Intra ETL Database Packages Well Defined Interfaces Sample Reports & Dashboard using OBIEE Sample IBM 4690 POS Adapter Oracle Retail Data Model Store Operations Point of Sale Loss Prevention Merchandising Inventory Subject Areas Workforce Management Order Management Example - Analytic Traffic Patterns, Comparative Store Performance, What Sells vs. Doesn t POS Flow, Shopper Conversion, Transaction Value, Time of day, Entry Method, Time Series, Trend Sweetheart Deals, Outliers for Return- Discount, Shrinkage, Employee-Basket Analysis, Trend Item-Basket, Product Stars & Dogs, Frequent Item Mix, GMROS, Cluster item Traits, Trend Sales Anomalies out-of-stock, zero selling, Forecast & Score, Time Series, Supplier Scorecard Measures AUS, AVS, UPT, Prescriptive Deployment, SPIFF and Split, Plan vs. Actual Channel Volume, Web commerce and interactions, Fulfillment Performance, Customer Order Analysis Customer Segment - Formation, Migration, Analysis, Price Elasticity, RFMP, Churn Model, Trend Database EE Category Management Promotion Assortment/Product Mix, Clustering, Plan-ogram, Customer Purchase vs. Syndicated data, Trend Causal Factor, Lift, Halo Impact, Predictive Response Model, Predictive cross-sell 11
Oracle Retail Business Intelligence Space Major Components of the application / technical stack Oracle Retail Data Model Oracle Retail Analytics Oracle Retail Workspace Oracle Retail Planning Applications Oracle Performance Management Applications Oracle ERP & CRM Business Intelligence Applications Oracle Business Intelligence Technology Oracle Retail Operational Applications Oracle ERP & CRM Operational Applications Oracle Database Oracle Fusion Middleware Oracle Database Appliance Machine - organize and discover data Oracle Big Data Appliance Machine - stream and acquire data Oracle Analytics Appliance Machine - visualize, analyze and decide data 12
In-Database Analytics Complete Oracle Data Warehouse / BI Solution Oracle Big Data Appliance Optimized for Hadoop, R, and NoSQL Processing Oracle Big Data Connectors Oracle Exadata System of Record Optimized for DW/OLTP Oracle Exalytics Optimized for Analytics & In-Memory Workloads Hadoop Open Source R Oracle NoSQL Database Oracle Event Processor Oracle Big Data Connectors Oracle Data Integrator Oracle Advanced Analytics Retail Data Model Oracle Database Oracle BI Applications Oracle BI Tools Oracle Enterprise Performance Management Oracle Endeca Information Discovery Oracle Real-time Decisions Times Ten Stream Acquire Organize Discover & Analyze 13
Data Collection and Transformation ORDM Technical Architecture Data Sources Point of Sale E-Commerce Partner Channels Merchandising Loyalty/CRM Financial ERP Allocations Social Media Competitive Supply Chain Derived Tables Foundation Layer Analytic Layer Presentation Layer Exalogic Big Data Appliance Exadata Exalogic Exalytics 14
ORDM Components Relationship Relationship between each component of the ORDM product Oracle Retail Data Model Source Data - OLTP System - Data Marts 15 - MDM, etc. Landing Zone - Tables / Views - CSV Files, etc. Staging Area - Data Quality - ETL Rules - Interface, etc. ORDM Foundation & Reporting Layers Master data is stored in Reference and Lookup tables Base tables stores only transactional data (3NF) Transactional Reporting Fact data is stored in Derived and Aggregated tables Analytical Reporting
Sample Analytical Report (Forecasting) Sales Trends vs Stock predict Out of Stock Demonstrates predictive analysis on sales forecast and inventory stock Oracle Retail Data Model provides many embedded forecast algorithms Oracle Exadata provides extreme performance for daily POS transactions 16
Sample Analytical Report (Product Category Mix Analysis) 17 Shows Market Basket Analysis using Key Performance Indicators (KPIs) Oracle Retail Data Model provides flexibility to identify correlations and their strength Report contains additional qualifying Basket/Component KPIs to identify interesting / useful rules Oracle Exadata provides extreme performance for ultra-fast cross sell analysis
Sample Analytical Report (Price Elasticity Analysis) 18 Retail: Price Elasticity helps determine the effect of applying a discount on a particular Item/SKU and analyze the impact on the bottom line (Revenue) This report allows the analyst to interact with the Mining Model via a dynamic application of the Discount % Can fine tune the discount % (not just steps of 1 but arbitrary value keyed in textbox by the analyst) Can apply it to a specific product and interactively see the impact on Revenue
Sample Data Mining Packages and Data Model Types Model Model ETL Package Model Creation Package Rule Associate Basket Analysis Model PKG_POP_DM_ASSBAS PKG_RBIW_DM_ASSBAS ABN, DT Associate Loss Analysis Model PKG_POP_DM_ASSLOSS PKG_RBIW_DM_ASSLOSS ABN, DT Associate Sales Analysis Model PKG_POP_DM_ASSSLS PKG_RBIW_DM_ASSSLS ABN, DT Customer Category Mix Analysis Model PKG_POP_DM_CUSTCATGMIX PKG_RBIW_DM_CUSTCATGMIX APASS Customer Loyalty Analysis Model PKG_POP_DM_CUSTLTY PKG_RBIW_DM_CUSTLTY ABN, DT Frequent Shopper Category Mix Analysis Model PKG_POP_DM_FSCATGMIX PKG_RBIW_DM_FSCATGMIX APASS Item Basket Analysis Model PKG_POP_DM_ITMBAS PKG_RBIW_DM_ITMBAS ABN, DT Item POS Loss Analysis Model PKG_POP_DM_ITMPOSLOSS PKG_RBIW_DM_ITMPOSLOSS ABN, DT POS Flow Analysis Model PKG_POP_DM_POSFLOW PKG_RBIW_DM_POSFLOW ABN, DT Store Loss Analysis Model PKG_POP_DM_STRLOSS PKG_RBIW_DM_STRLOSS ABN, DT ABN = Adaptive Bayes Network DT = Decision Tree A = Apass (Market Basket Analysis) 19
Sample Analytical Report (Customer Loyalty Analysis) Identifies attributes that have significance in predicting loyalty Segment Customers and determine loyalty Can apply findings to identify prospects who fit the loyalty profile Oracle Exadata quickly finds transactions of customers in a given loyalty category 20
Q & A 21