"Data governance is currently a relatively new and developing discipline, and its definition in the industry is not exactly the same at present." Xie Guozhong, deputy partner of the business analysis and optimization team of IBM Global Enterprise Consulting Service Department, said that data governance is a series of specific work around taking data as enterprise assets. Data is the biggest source of value and risk for enterprises. Poor data management usually means poor business decisions and is more likely to face violations and theft. Using the trusted data of rules is helpful to the business innovation of organizations-providing better services, improving customer loyalty, reducing the work required for compliance and reporting requirements, and improving innovation ability.
The data governance maturity of domestic enterprises is not high.
In the past few years, the goal of data governance has also changed. Xie Guozhong, who once helped customers in financial, aviation, customs, telecommunications, power grid and other industries to do data governance projects, said, "In addition to meeting the requirements of supervision and risk management, many enterprises are now talking about how to create business value through data governance. Such as information disclosure, industry leadership, and refined management needs. "
"At present, most domestic enterprises are still in the basic management stage in data governance," Xie Guozhong commented. "Some companies say that they have done a lot of data quality inspection, data archiving and data security, but their problem is that they don't have a complete system. Secondly, they don't know how to string these fields together. Third, they have not yet reached the concept of taking operational data as a core asset. " Therefore, Xie Guozhong believes that domestic enterprises need a complete data governance system first.
He believes that there is a misunderstanding in data governance of domestic enterprises: data governance is a very short-term behavior, and data governance is only the responsibility of IT departments, and data governance is only regarded as software. In fact, data governance is not only software, but also corresponding processes and methods.
When it comes to the practice of data governance, IBM itself is a typical representative of data governance. Before 1992, IBM had many problems in data governance, such as no clear and reliable data source, no clear data owner and low data quality. 1995, IBM established business data standards in ERP, and established 15 business standards and 79 sub-business standards for all businesses, which made the whole company see a unified business definition. In 2004, IBM set up a data owner forum, a data governance committee in 2005, and then a data audit committee. In 1992, IBM had 128 chief information officers, 155 data centers, 80 Web development centers, 3 1 different networks and 16000 applications. Through data governance, IBM simplifies the infrastructure and reduces the complexity of management. In 2007, IBM had only one CIO, six major data centers, a global unified network for Web development centers, and about 4,000 applications from 65,438+06,000.
On this basis, in 2004, IBM established a data governance forum in conjunction with many companies and academic research institutions in the industry, and made a data governance framework and method including four major areas 1 1 on this forum to guide the development of data governance. The framework includes output fields: data risk management and value creation; Driving fields: organization/process, management system, data owner; Core areas: data quality management, information life cycle management, security/information disclosure/compliance; Support areas: data model/data architecture, metadata/master data/data standards, quality audit and reporting.
Successful cases of bank data governance
In terms of data governance, due to the promotion of policies and the needs of the bank's own business development, banks have a strong demand for data governance. The 12th Five-Year Plan of China Banking Information Technology contains the theme of data governance and data standards, and points out that the core areas that need to be promoted in data governance during the 12th Five-Year Plan period include data standards, data quality, data security, data architecture, and the guarantee mechanisms needed to do these tasks well, including policies, organizations, processes and technologies.
IBM GBS department has helped domestic and foreign banks to do many data governance consulting projects, including the largest commercial bank in China.
"This bank is the most technologically advanced in the same industry, and has done data quality and metadata, but there is no complete and unified data governance method and supporting systems and processes; Another problem is that the data governance system and architecture are still not perfect. " Xie Guozhong briefed reporters on the challenges faced by the bank in data governance.
According to IBM's data governance framework and method, GBS evaluates the status quo of bank's data governance from the 1 1 elements in four major fields, helps banks find the gaps, and analyzes the problems on this basis, and puts forward suggestions to solve the problems. The project started in 2008. During the periods of 20 10 and 20 1 1, the bank successively carried out a series of related data governance projects, including data standardization projects, data architecture optimization projects and data quality management projects. At present, the bank is also in the forefront of the country in terms of data governance.
Take a look at the data governance case of a leading global bank. The Fed believes that the bank did not fully control the integrity and quality of information to ensure compliance requirements. After one or two years of data management, the bank passed the audit of the Federal Reserve. Andrew Dunn, senior vice president of the bank, believes that the key factor for the success of data governance in the bank is that choosing a partner with relevant experience, processes and tools can accelerate the effective deployment of data governance throughout the enterprise.
The demand for master data management is very prominent.
Data governance involves 1 1 element, and master data management is a very important part of it. Tony Young, senior vice president and chief information officer of Informatica, said: "Strengthening master data management is the only way for enterprises to obtain a complete and credible data view."
Master data is used to describe the core business entities of an enterprise, such as customers, partners, employees, products, bills of materials, etc. Master data management aims to integrate the core data that needs to be shared in multiple business systems of an enterprise, centrally clean up the data, and distribute the unified, complete and accurate master data to the operation and analysis applications of the enterprise in the form of services, including business systems, business processes and decision support systems.
Tony Young told reporters: "The core task of MDM is to output' golden data'. The so-called golden data is the key business data of the enterprise and the absolutely true data. In addition, MDM should also reflect the correlation between master data, such as the relationship between customers and products, and the relationship between customers. In the data warehouse, it is difficult to find this correlation, and MDM can easily do this. " There are also differences between MDM and data warehouse. For example, they handle different types of data. MDM is a transaction system, and data warehouse is an analysis system. MDM and data warehouse can promote and complement each other. Informatica MDM's flexible data model allows IT teams to implement MDM in any data domain, and can add other domains in the same data model and define the relationships between different data domains. Informatica MDM can be implemented within the enterprise or in the cloud, or as a mixture of the two. In addition, it can also be deployed in a federated MDM architecture as a global hub between multiple MDM instances.
The financial industry is still the industry with the strongest demand for MDM. Andy Hayler, a senior expert of MDM and president of Information Difference, said: "Generally speaking, the bigger the enterprise, the more problems it will encounter in data management. Large companies are more likely to adopt relevant data analysis tools to solve the data problems they face. " But this does not mean that MDM is only applicable to the financial industry. For example, Informatica's MDM products have been applied in 24 industries, including medical, petroleum, public utilities and other industries.