Master Data Management Detailed

What is master Data management (Master Management, MDM)

Master data is the data to be shared across the entire enterprise system (operational/transactional applications and analytical systems), e.g., with customers (customers), suppliers (suppliers), accounts (accounts), and organizational units ( Organizational units) related data. Master data typically needs to be consistent across the enterprise (consistent), integrity (complete), controllability (controlled), and in order to achieve that goal, master Data management (master database Management, MDM) is required. It should be noted that the main data is not all business data in the enterprise, it is only necessary to share the data between the various systems is the main data, such as most of the transaction data, billing data, and so is not the main data, but like the core business entity data, and like customers, vendors, accounts, organizational units, employees, partners, bit Information, etc. are master data. Master data is high value data that can be reused across business in an enterprise. These master data often exist in multiple heterogeneous or homogeneous systems prior to master data management.

Master Data Management (Master Management, MDM) is a set of constraints and methods used to ensure the real-time, meaning, and quality of data related to the subject domain and systems within an enterprise, as well as across subject domains and systems. This is a profound illustration of the depth and complexity of active master Data Management (MDM), and simply, Master Data Management (MDM) ensures that your system coordinates and reuses common, correct business data (master data). Typically, we add master data management as a complement to the application process, by separating the main information from various operational/transactional applications and analytical applications, it becomes a centralized and independent core of other applications in the enterprise, thus enabling the Enterprise's core information to be reused and ensuring that all operational/transactional applications as well as Consistency of core data between analytical applications. Through the main data management, change the status of enterprise data utilization, so as to better pave the way for enterprise information integration.

Master Data Management (MDM) helps us create and maintain a single view of master data across the enterprise, ensuring the accuracy, consistency, and integrity of a single view, providing data quality, unifying the definition of business entities, simplifying business processes, and delivering business responsiveness. From the frequency of change, the main data and daily transaction data is not the same, the change is relatively slow, in addition, the main data due to across the system, so the data consistency, real-time and version control requirements are high.

Master data management actually existed long ago, just now with the business development and regulatory needs, the main data real-time, accuracy, consistency has a higher demand, only to be widely accepted by the industry, each manufacturer accordingly launched a series of master data Management integration and basic suite as well as specific areas of the solution. The most obvious change in recent years has been the question that customers often ask in the past: "What is master Data management?", and now customers often ask questions evolve: "Our business does have some problems, master data management can solve this problem, how do we start?". Customer awareness of Master Data Management (MDM) has improved dramatically compared to the previous and start experimenting with master Data Management (MDM) to address the challenges and problems that they encounter across business, across the topic domain throughout the enterprise: for example, in the tax industry, the Inland Revenue Department, according to the taxpayers in some statistical analysis, It is found that the basic information on the taxpayer is distributed in the core collection management system, invoice management system, personal income tax system, VAT management system and so on up to dozens of systems, making statistical analysis difficult, in the medical equipment companies, for example, due to the lack of product classification according to the supplier, The description of each product is also very different, making the product catalog maintenance very difficult. With the development of the business, it is very urgent and necessary to build and maintain a unified master Data system for all walks of life, especially for some TNCs, it is important to maintain a single view of information about customers, product catalogs, suppliers, etc. between the different regions (countries and regions) business systems.

It should be noted that master data (master) and metadata (meta data) are two completely different concepts. Metadata refers to information related to data, such as data definition, and master data refers to instance data, such as product catalog information. For example, a provincial government tax has developed a collection of management software, to the city as a unit deployed 17 sets, each collection of management software metadata are the same, but the main data still need to be managed. Master data management and traditional data warehouse solutions are not a concept, the Data Warehouse will centralize the data of each business system in the business analysis, while the master data Management system does not manage all the data, but the need to share among the main data collection and distribution. Compared to the one-way integration of traditional data warehouse solutions, master data management is focusing on synchronizing changes in master data to each associated business system (master Data Management data is bidirectional).

The root of the problem of master data management

For most businesses, there is a problem with master data management, personally think this is due to business development and the gradual development of it technology, it is precisely because of this progressive, the business system of major enterprises from scratch, from simple to complex, thus forming one after another of the business shaft. Fundamentally speaking, it is impossible to use only one business system to cover all the business of the enterprise, even for the packages provided by some international large companies, which is an impossible task (even for a suite, there is often a multinational enterprise deploying multiple instances in different countries or regions, i.e. no centralized Department of the kit, but in many places dispersed deployment of the kit). For the enterprise, the construction of the business system is more project-centric, Bottom-up's construction system, rather than the supremacy of the building system, is bound to lack the entire enterprise-wide unified planning, so that some need to share in the various business data (master data) are dispersed to the various business systems for separate management. Due to the lack of consistency, accuracy and completeness of the main data of decentralized management, there are many phenomena, such as ineffective product management, ineffective supplier management and poor order management. The fundamental way to address this problem is to introduce master Data management (MDM), which not only refers to data that needs to be shared, but also business rules and policies that need to be shared.

The maturity of master Data Management (MDM)

Based on the complexity of the implementation of master data management, according to Jill Dyche, Evan Levy's view can roughly divide master data management into five levels, from low to high, reflecting the different maturity of master data Management (MDM). Here's a brief introduction to these five levels:

Level 0: No master Data Management (MDM) is implemented

In the case of level 0, it means that there is no data sharing between the various applications of the enterprise, and that no data definition element exists across the enterprise. For example, a company sells many products, the production and sales of these products are handled by a number of independent systems, each system handles the product data independently and has its own independent product list, each system does not share the product data. At level 0, each independent application manages and maintains its own critical data (such as product lists, customer information, and so on), which are not shared between systems and are not connected.

Level 1: Provide a list

Whether the company is large or small, list management is one of our common ways. Within the company, a logical or physical list is maintained by manual means. The list is available when heterogeneous systems and users need some data. The maintenance of this list, including data additions, deletions, updates, and conflict handling, is handled through a series of discussions and meetings by the staff of each department. Business rules (Business rules) are used to reflect the consistency of values, and when business rules change or a similar situation occurs, this highly manual-managed process is prone to errors. Because list management is managed by hand, the quality of its list maintenance depends on who participates in the change management process, and the maintenance of the list will be affected once someone is absent.

The difference between MDM Level 1 and MDM level 0 is that each department maintains its own critical data independently, but maintains a loose list of master data through list management, providing the data it needs to other departments. In MDM Level 1, data change decisions and data change operations are determined by people, so only people who have completed data change decisions will change the data. In fact, although the data change process has strict rules, but because of the lack of centralized, rule-based data management, when the data volume is large, the cost of data maintenance will become very high, efficiency will be very low. List management is possible when the number of master data, such as customer information, product catalog information, etc. is relatively small, but the change process of list management becomes difficult when the product catalog or customer list explodes. MDM Level 1 relies on human collaboration. If the product manager needs to update the product price list later, it needs to contact the ERP system owner to send it to her. Implementing a customer or product list on an enterprise-wide scale is just like maintaining relationships among people in different departments. If there is a hierarchy or grouping of customers or products, the list will be difficult to provide, and usually at level 1 because it is too complex to manage.