SPECIAL CONFERENCE ISSUE THE OFFICIAL PUBLICATION OF THE Orace Appications USERS GROUP spring 2012 Introduction to Master Data and Master Data Management (MDM): Part 1 Utiizing Orace Upgrade Advisor for R12 Best Practices for Push & Pu Using Orace Inventory Stock Locators Networking: The Power Boost for Your COLLABORATE 12 Experience!
oaug feature This artice is the first of a two-part series. Part 1 provides an overview of what Master Data is and discusses the chaenges associated with managing Master Data. Part 2 concudes the artice with a ook at the art and science of Master Data Management and shares insights into the atest innovations in process and technoogies. Part 2 wi appear in the summer 2012 OAUG Insight magazine. Introduction to Master Data and Master Data Management (MDM) (Part 1 of 2) By Mani Kumar Manda, Rhapsody Technoogies, Inc. As Master Data Management (MDM) appications mature to the eve of suitabiity for enterprise depoyment, many companies wi want to competitivey differentiate themseves with a strategic impementation of MDM soutions. Every day, companies ose miions of doars due to making business decisions based on inaccurate Master Data. The abiity to use data that is free of dupications and errors when making critica business decisions and conducting marketing campaigns is of inestimabe vaue. This artice defines Master Data, addresses the chaenges of managing Master Data and shares insights into the atest innovations in process and technoogies. 2 OAUG FEATURE 16
This artice appeared in the spring 2012 issue of OAUG Insight magazine, the officia pubication of the Orace Appications Users Group (OAUG), and is reprinted with permission. oaug insight spring 2012 Data Spectrum Per the Merriam-Webster Dictionary, the word data is defined as factua information (measurements or statistics) used as a basis for reasoning, discussion or cacuation. The term digita data is used to refer to the capture, storage and movement of data using digita technoogies by encoding it in numerica forms that can be easiy transmitted or processed. Digita data forms the foundation for modern day commerce and can be broady cassified into two categories: structured and unstructured data. For the purpose of this artice, the terms data and digita data are used interchangeaby. Digita data is growing exponentiay, with growth chaenges being three dimensiona: 1) more data is being created (amount of data), 2) data is being created faster and needs to be more quicky accessed (the veocity) and, asty, 3) the types of data being captured is broadening (variety). The vast amounts of data being created in the digita media are unstructured and require technoogies that are specific to processing and managing it. The further discussion about unstructured data is outside the scope of this artice. Structured data, which is addressed by MDM among other technoogy soutions, represents a esser percentage of a data being created and is managed through technoogies that are aready we deveoped. This incudes reationa database technoogies that can easiy store, access and process such data. Structured data is cassified into the foowing categories (aso shown in the iustration to the right): Metadata (Data about Data) Structured Data Reference Data refers to the types of data that are fairy static in nature but do change over a ong period of time. Reference data fas into one of four categories: a) data used as a cassification scheme such as Standard Industry Cassification (SIC) codes.; b) codes that are externa to the business organization, such as country codes, currency codes, etc.; c) ookup codes such as status code using the vaues of active, inactive; and d) constant vaues such as tax rates. Reference data requires itte effort to maintain, athough the primary chaenge in maintaining it is to ensure the accuracy of mapping across heterogeneous appications and incorporating it in a integrations in a timey fashion when new reference data is created or existing reference data is changed. STRUCTURED DATA Reference Data Master Data (Nouns) Unstructured Data Transactiona Data (Verbs) Audit Data Archive Data Reference Data Master Data Anaytica Data (Facts & Dimensions) Transactiona Data Anaytica Data Metadata Reference Data Exampes: Lookup Codes Master Data Exampes: Customer, Product Audit Data Archive Data Transactiona Data Exampes: Orders, Invoices, Leads Anaytica Data Exampes: Facts, Cubes, Dimensions oaug.org 17
Master Data is dynamic. It changes continuousy, with or without the invovement of an organization... Master Data is simiar to Reference Data, but changes graduay over time, due to both externa and interna factors to an organization. Master Data usuay refers to nouns such as customers, empoyees, products, suppiers, etc. Organizations need to devote resources, both in peope and technoogy, with the hep of wedefined processes in order to keep this data at a consistent and accurate eve for organizationa users to consume it effectivey. Master Data uses Reference Data. Transactiona Data, commony associated with verbs and considered the record of an event, refers to data that captures information about a specific event such as an order, an invoice, a ead, etc. Transactiona Data uses both Master Data and Reference Data. Anaytica Data refers to data structures that are optimized for anaytica purposes to infer knowedge based on massive amounts of data in shorter periods of time. Data for anaytica purposes is stored into facts and dimensions, with data modes foowing principes of star schema or snow fake schema to meet performance objectives. Anaytica Data uses a three of the data types defined above: Transactiona Data, Master Data and Reference Data. Metadata refers to the data that describes other data. Audit Data refers to data that describes changes made to existing data, such as who performed the change, when the change was made and in which system the change occurred. Audit data is captured and stored for the purpose of capturing data changes and to prevent or identify root causes of data changes when a data change impacts the business in some form and fashion. Archive Data refers to the storage of a structured data that is presenty not needed but is required to be accessibe for the purpose of compiance with audit and reguatory reasons. Data is aso archived in order to improve the efficiency in terms of performance of operationa appications. How ong data is archived varies greaty by industry and ranges anywhere from seven to more than 30 years for firms in the insurance industry. Let s take a coser ook at Master Data. Master Data Defined Master Data, previousy defined as nouns, refers to the type of data that has a ong ife span and changes graduay over time. Master Data can be cassified into four categories, as shown in the chart beow. Each type of entity under these categories is known as a domain. Master Data Categories & Domains (or Entities) Parties things Paces concepts Exampes: Customer Suppier Empoyee Citizen Partner Bank Exampes: Product Part SKU Service Asset Exampes: Faciity Location Geography Exampes: Chart of Account Contract License Poicy 18 O AUG FEATURE
oaug insight spring 2012 Parties refer to entities such as individuas and organizations that are invoved as trading partners in conducting business. The Parties can be empoyees, customers, suppiers, citizens, etc. The management of customer data is known as Customer Data Integration (CDI). Things refer to items such as products, parts, SKUs, etc., that are used in commerce. The management of product data is known as Product Information Management (PIM). Paces refer to data entities such as faciities, ocations and structures that are tied to a physica address in a geographic space and the maintenance of a their associated information. Concepts refer to data entities such as chart of accounts and contracts that are a means to group/ cassify to encapsuate pertinent data. For exampe, chart of accounts are used to group journa entries as per accounting principes. Contracts are used to capture a business reationship in contractua terms that govern the business interaction. What Master Data domains are important to address and the reative criticaity of each domain to an organization depends on the industry they are in and the types of products and/or services these organizations offer. For exampe, for an organization that buids and ses airpanes, given the number of parts and subassembies used in each of these panes (often gets into miion pus parts) and the revisions that are made to these panes and the subassembies used over the years, the soutions to address both product Master Data and suppier data become more critica. Whereas, for distribution companies and organizations that operate in the B2C (consumer) space, the focus on customer data is more important, with some of them aso needing soutions to address product Master Data. The Chaenge Master Data is dynamic. It changes continuousy, with or without the invovement of an organization, and is universa across a domains of Master Data beonging to the four categories discussed. According to Dun & Bradstreet, 17 percent of business names change every year, resuting in changes to customers, suppier and a other domains in which a business pays a roe. Per the United States Posta Service (USPS), about 14 percent of the popuation moves yeary, resuting in address changes for suppiers, customers, empoyees, citizens, etc. Facts such as these indicate the dynamic nature of customer data. Most of these changes often are not communicated to organizations. Businesses are, therefore, eft aone to devise their own soutions and approaches to obtain the most accurate data, sometimes using sources other than the customer itsef. Many other chaenges are tied to the management of party data. According to Dun & Bradstreet, 17 percent of business names change every year, resuting in changes to customers, suppier and a other domains in which a business pays a roe. Facts such as these indicate the dynamic nature of customer data. The nature of product (things) data is such that most products are uniquey identified by their characteristics (which often form the basis for generating item descriptions). Though a unique identifier (item number) makes it appear that there are no dupicate products in the database, the reaity is often different. It is frequenty discovered that two or more products oaug.org 19
defined with unique identifiers indeed are the same products. And mutipe products are often created by different users, with the characteristics represented in the description in a unique manner (descriptions created with representation of seected characteristics in different order and/or using varying abbreviations) making it difficut to identify these products as dupicates. This scenario makes it very difficut to identify dupicates since the very nature of identifying dupicates cannot be easiy done using pattern matching technoogies. Identifying dupicates in product data requires a technoogy different from pattern matching, and semantics-based matching technoogies had great success in matching products for many industries. In essence Master Data is constanty (though sowy) changing. One estimate predicts that a perfecty accurate repository of Master Data becomes obsoete by at east 20 percent for every year it is not maintained. mutitude of appications then comparing the data where simiar information exists in mutipe appications and identifying the correct vaue when conficts exist. This is easier said than done. The reaity is that changes made in one appication often are not refected in other IT appications. The task of operating the business, which is based argey on insights gained from information for which Master Data forms the basis, is error prone because Master Data is often not uniform across the organization. In essence Master Data is constanty (though sowy) changing. One estimate predicts that a perfecty accurate repository of Master Data becomes obsoete by at east 20 percent for every year it is not maintained. Other measures peg the change in Master Data at much higher rates, even approaching two percent per month. Compounded, this monthy change becomes a change of 27 percent of data in one year, a change of 64 percent in two years and a whopping 104 percent change over three years. What shoud organizations do to address these chaenges? Watch for the concusion of this artice in the summer 2012 issue of OAUG Insight magazine. Simiar chaenges exist for other domains of Master Data, incuding data in the ream of sites or concepts categories. Another chaenge with Master Data is that it is often scattered across the organization, housed in various appications that are used by a specific business ine, business function, purpose or geography. It woud take a tremendous effort to compie a fu portrait of the Master Data of an instance, e.g., specific customer, specific suppier, specific product, specific site, etc. This task requires first matching simiar data across a Mani Kumar Manda is founder of Rhapsody Technoogies, Inc. and founder/coordinator for the OAUG Customer Data Management Specia Interest Group (OAUG CDM SIG). Mani has been a consutant for his entire career to mid-market and arge customers and is a recognized Expert and Speaker in the areas of MDM with numerous presentations at OAUG, Open Word, MDM Summit and severa oca GEO and SIG groups. Mani has deveoped a proprietary RHYTHM Methodoogy that significanty increases the success factor for a MDM impementations. Mani can be reached at mmanda@rhaptech.com. Acknowedgments: Specia thanks goes to Eugene Breger for reviewing and providing changes that enhanced the quaity of this version of the artice. 20 OAUG feature
oaug insight spring 2012 The reaity is that changes made in one appication often are not refected in other IT appications. References 1. Customer Modeing: It s your ca between Rock and Hard Pace, http://www.sideshare.net/ RhapsodyTechnoogies/customer-modeing-its-your-ca-between-rock-and-hard-pace. 2. Orace CDH the past (R11i), the present (R12), and the Future (Fusion MDM), http://www.sideshare. net/rhapsodytechnoogies/orace-cdh-the-past-11i-the-present-r12-and-the-future-fusion. 3. Orace TCA 101 (with audio track), http://www.sideshare.net/rhapsodytechnoogies/orace-tca-101. 4. MDM Orace Site Hub 101, http://www.sideshare.net/rhapsodytechnoogies/mdm-orace-sitehub-101-1419759. 5. Site/Location Hubs A Hot Trend in Master Data Management (MDM), http://www.sideshare.net/ RhapsodyTechnoogies/siteocation-hubs-a-hot-trend-in-master-data-management-mdm. 6. R12 features and functionaity in TCA and CDH/EBS, http://www.sideshare.net/ RhapsodyTechnoogies/r12-features-and-functionaity-in-tca-and-cdhebs. 7. Orace Trading Community Best Practices Setting Up Customer and Prospect Data, Search Orace s support site with Note# 269124.1; or https://support.orace.com/cgi-bin/cr/getfie.cgi?p_ attid=269124.1:2. 8. To receive an Orace CDH Poster, send an emai to OraceMDMPoster at RhapTech DOT com. Websites and Discussion Forums OAUG CDM SIG Rhapsody Technoogies, Inc. Product Data Quaity Suppier Life Cyce Management (SLM) and Suppier Data Hub (SDH) Orace Fusion Appications http://groups.yahoo.com/group/cdmsig http://cdmsig.oaug.org http://www.rhaptech.com http://www.rhaptech.com/resources.htm http://tinyur.com/29xocwv http://tinyur.com/2v2yyct http://tinyur.com/3232emw oaug.org 21