On April 27, at the 2023 Data Governance New Practice Summit, Mr. Jiang Chunyu, vice chairman of the Big Data Technology Standard Promotion Committee, shared the value of data and the latest development trends with the theme of "Data Governance Development Trends".
The following is the transcript of Mr. Jiang Chunyu's speech. For the convenience of reading, the editor has made some word modification and text optimization.
good morning guys! It is a great honor to be invited to participate in the 2023 Data Governance New Practice Summit, and it is also a great honor to be the opening guest to share with you the latest development trend of domestic data governance.
Data has become a strategic resource
The field of "data" has been hot for many years. Around 2012, big data technology was just introduced into China, and the data technology boom started; by 2015, the country issued the "Action Outline for Promoting Big Data Development", which marked the rise of data to the national strategy; in 2016, the country Released the 13th Five-Year Plan for Big Data Development. For the first time, Big Data has an independent five-year development plan as an industry. In 2022, "Data Twenty" was released. Data is becoming a factor of production, and new systems and mechanisms need to be built to release its value. .
In June 2022, "The Central Committee of the Communist Party of China and the State Council Issued Opinions on Building a Data Basic System to Better Play the Role of Data Elements", referred to as "Data Twenty Articles", made a lot of descriptions on the future basic system construction of data elements, including the data property rights system , transaction circulation system, income distribution system, security governance system, etc. The core purpose of the "Opinions" is to improve the quantity and quality of data element supply, fully protect the rights of data processors to use data and obtain benefits, fully realize the value of data elements, and promote the sharing of digital economic development dividends by all people.
From the perspective of enterprises, many enterprises are transforming into "data-driven enterprises". Most of them have basically achieved online business and accumulated a large amount of data. The most urgent proposition now is how to build a "data-driven enterprise". , Let intelligent decision-making and data-driven culture go deep into the small cells of every enterprise. We believe that data-driven enterprises have six major characteristics, namely, comprehensive use of data, real-time analysis, data available at any time, flexible management and control, internal and external integration, and closed-loop operation.
The Three Phases of Data Governance
Data governance is the basis for realizing the construction of a data element market and a data-driven enterprise. Recently, Ms. Meng Wanzhou from Huawei mentioned in public that data governance is the foundation of digital transformation: only through scientific governance of data can the flow of data within the enterprise be meaningful, and new value can be created when data from different dimensions come together. . The purpose of data governance is to make data available, usable, and easy to use, release the value of data, and finally realize data-driven enterprise operations.
The development of data governance has probably gone through three stages.
The first stage is the information age.
At this stage, data has not been stored in a large amount and extensively. Data governance is mainly to help enterprises build IT systems in a standardized manner and enable information sharing between IT systems. This is a feature of the information age.
The second stage is the era of big data.
After 2010, we entered the era of big data. With the development of Moore's Law, hardware performance has been continuously improved, enabling us to have cheap massive data storage and computing capabilities. At the same time, the mobile Internet has brought massive user behavior data, which can be stored , calculation and mining, applied to Internet advertising, personalized recommendation, intelligent risk control and other fields. Only when the data has commercial value can we enter the era of big data. In the era of big data, data governance faces challenges such as large scale, high timeliness, and complex data types, requiring automated, agile, and diverse data governance capabilities. The management, insight, and mining of comprehensive data assets are the characteristics of the big data era.
The third stage is the era of data elements.
After 2020, we are evolving into the era of data elements, the logic of which is that data is widely circulated and interconnected. Data not only needs to flow within the enterprise, but also between different enterprises. The demand for this flow is universal. of. The data governance system in the era of data elements is also different from the previous ones, because it is not only an issue of data governance within an enterprise, but also involves data governance issues between different organizations. If any of these subjects has shortcomings in security, compliance, and governance, then the data will flow away. Therefore, in the era of data elements, the challenge of data governance lies in the demand for multi-subject data flow.
Regarding the theory of data governance, from the earliest DMBook to DMM maturity to the domestic data management capability maturity model, various knowledge frameworks are constantly emerging. DMBook is partial to the knowledge system, which is a must-learn course for everyone entering the data governance industry; and the DMM maturity model draws on the form of the software maturity model to evaluate the data management capabilities of enterprises in the practice process.
Five trends in data governance
Next, I will focus on data governance trends. I have summarized the following five trends:
Development trend 1: DCMM implements standards to promote data governance capacity building in various industries
The first trend is that we see that domestic data governance methodologies are taking shape. DCMM is the first national standard in the field of data management in my country. It was established in 2014 and released in 2018. In 2020, it will start to carry out standard implementation evaluation work nationwide. The DCMM standard defines 8 capability domains, 28 capability sub-domains, 445 items, and 5 capability levels. The promotion of DCMM at the national level is sparing no effort. The "Twenty Data Measures" proposes to accelerate the implementation of the national standards for the maturity of data management capabilities and the management of data elements. The Ministry of Industry and Information Technology's Big Data Industry Plan (2021-2025) also proposes to improve the data management capability assessment system, promote the implementation of national standards such as the "Data Management Capability Maturity Assessment Model" and data security management, and continuously improve the data management level of enterprises and institutions.
2022 is the first year of explosive growth in China's data governance. In 2022, various evaluation agencies have completed the DCMM standard implementation evaluation work of more than 1,000 enterprises. With the strong support of various local governments, the concept of data management and data governance has been implemented in all walks of life. The awareness of DCMM has been greatly improved, and the demand for data governance has become hotter.
Development trend 2: Data governance urgently needs to be integrated into data development
We have seen that leading enterprises such as large banks and large operators have built relatively powerful data governance systems, but there are generally two problems of data governance and data development. How to embed governance capabilities into the data development process, Accelerating the efficiency of data development, breaking down the barriers of collaboration among teams, and forming a data production pipeline integrating data development and governance have become urgent needs of leading enterprises.
We are exploring the use of DataOps (integration of data R&D and operation) to build an efficient collaboration mechanism, create an integrated development and management pipeline, and establish a refined data operation system. DataOps is a new paradigm of data development. It integrates concepts such as agile and lean into data development. By reorganizing data-related personnel, tools, and processes, it breaks down collaboration barriers and builds automated data that integrates development, governance, and operations. Pipeline, continuously improve the efficiency and quality of data product delivery, and achieve high-quality digital development. This is our definition of it.
The Institute of Information and Communications Technology will gradually promote the practice and implementation of the concept of DataOps in China. The pipeline of DataOps is divided into four domains, namely R&D, delivery, operation and maintenance, and value. It has three key security functions, namely organizational management, system tools, and security control. The figure below shows a framework diagram of Dataops. We are promoting this framework to be implemented in various enterprises. Leading organizations have begun to consciously use DataOps ideas to solve the problem of insufficient data productivity.
The DataOps standard and practice system is taking shape. By the end of 2022, we have completed the first standard: the standard system for R&D management, and then implemented it in the Agricultural Bank of China, ICBC, Zhejiang Mobile, and Jiangsu Mobile. In 2023, we plan to complete the delivery, operation The development of three standards of dimension and technical tools; the two standards of value operation and organizational management will be completed in 2024. In the end, seven standards will be formed, and we will gradually improve the standard system and gradually promote the implementation of the concept of DataOps in China.
Development trend 3: Data governance leaps forward to data asset management
It is far from enough to talk about data governance. We have learned that some companies with better data foundations have begun to explore the work of data capitalization. We have proposed the concept of data assets since 2017, and promoted the development of data assets in China. The "White Paper on Data Asset Management Practice" we released has been updated to version 6.0. Data will not only generate a lot of problems, but also generate a lot of value. How to measure and reflect these values inside and outside the enterprise? Two stages of data resourceization and data assetization are mentioned here. Much of the data governance work we do now is to manage raw data as data resources, so the next step is to release the value of data, that is, to capitalize data. Data assetization has three major activities: asset valuation, operation, and circulation, which are also defined in the white paper.
Regarding valuation and operation, we have done some related work. We have seen that many institutions in the industry are actually exploring valuation, most of which are based on Gartner’s information value evaluation model. There are three evaluation methods: revenue, cost, and market. These three evaluation methods actually have some problems. We believe that the valuation of data assets needs a framework and some core indicators, and we are defining the valuation framework and indicators. There are two valuation methods: monetary method and demonetization. We believe that it is more realistic for companies to conduct non-monetary value evaluation around digital scenarios. Of course, after completing non-monetary evaluation and value measurement, monetization conversion can be carried out in the future. This is a two-step relationship. In addition, we are cooperating with central state-owned enterprises on some corresponding topics. Of course, their data foundation is relatively good, basically gathering data from the whole domain, the data quality is good, and the data application scenarios are more. Therefore, we are sorting out the value of data in the scenario based on the data application scenario.
Leading companies aggregate data and generate a lot of value on the business side, but this value is far from enough. Because businesses do not have a deep understanding of data, we introduced the concept of data asset operation. The data team must take the initiative and explain to the business what value it has. Data operation includes five stages: asset planning, asset introduction, asset promotion, asset use, and asset optimization. Nowadays, the data teams of enterprises generally have to do a lot of work, including technical work that has gone beyond the data itself. There is no other reason, because it is difficult for the business to understand the data content. In order to achieve results, the data team must do more. Although the ultimate goal is to allow the business to use and retrieve the data, there must be a process in the middle. That is, the data team must first embrace the business. It not only needs to organize, aggregate, and process the data well, but also needs to tell the business how to use the data and better help the business to release the value of the data.
Development Trend 4: There is an urgent need for data security governance construction
The national "Data Security Law" requires enterprises to build a data security governance system to deal with various data security risks. This requirement is generally applicable to advanced enterprises with a high level of data governance, that is to say, enterprises above the designated size must build systematic data security governance capabilities. We are also launching a data security governance capability assessment framework. This assessment framework not only defines the concept and framework of data security governance, but also tells everyone how to build a data security governance system step by step.
At present, we have taken the lead in landing in advanced enterprises including Ping An Bank, Shanghai Pudong Development Bank, Baidu, and Ant. During this process, we found that these enterprises have been doing data security-related work for at least five years. Enterprises should pay attention to data security governance. They need to comprehensively evaluate the overall capabilities of data security from security planning, life cycle, and basic security, and implement it through a systematic methodology.
In addition, the construction path of data security governance is gradually becoming clear. In the "Data Security Governance Practice Guide 2.0", we put forward the practice path of "planning-construction-operation-optimization". Governance planning can be used for status analysis, program planning, and program demonstration; governance construction is based on data life cycle scenarios and business operation scenarios to sort out and implement. Governance operation requires risk prevention, internal control early warning and emergency response; governance optimization refers to continuous internal and external evaluation and optimization.
Development Trend 5: Intelligent Data Governance
The intelligence of data governance is also a very important development trend, especially the new generation of artificial intelligence technologies represented by ChatGPT and graph intelligence, which can play a huge role in data governance. First, it can improve the efficiency of data governance and reduce a lot of manual work, such as classification and grading; second, it can enhance metadata management capabilities, and improve blood relationship analysis and impact analysis capabilities.
In addition, the impact of ChatGPT on the entire data governance, in the short term, I judge that it is good for data governance, because it requires a large amount of data corpus, and data governance happens to be a necessary part of improving data corpus.
Summarize
Finally, to sum up, data governance is a prerequisite for the release of data value.
Data governance is facing five development trends.
Trend 1: Our domestic methodology has been gradually formed, and DCMM evaluation is used to promote the construction of the entire data governance system, especially for large enterprises, which use DCMM to attract the attention of upper-level leaders.
Trend 2: Data governance and data development should be integrated, they cannot be two skins, otherwise the overall efficiency and application value will be compromised. DataOps is an innovation in the data production model. Various industries have certain practices, and standardization work is in progress.
Trend 3: Data asset management is the stage of data governance 2.0, and data valuation and value operation have become the "compass" of digital transformation.
Trend 4: Data security has become the basic bottom line of data utilization. Systematic data security governance capabilities are very necessary and a guarantee for our enterprises to deal with various data risks.
The last trend is intelligence, which improves the efficiency of data governance and promotes the value of data governance. It is a very important trend in the evolution of data governance in the future.
The above is the sharing of all data governance trends! Thanks for listening!