Today a digital wave driven by big data, the Internet of Things and artificial intelligence is sweeping the world. Huge volumes of data are being generated at every moment, and once the quantity and scale of that data reach a certain level, existing storage, analytics and computing solutions and technologies will no longer meet real-world needs. A company's existing big data is like a gold mine, and its high-value data assets are the gold waiting to be mined: only after a series of processing steps can the value within them be extracted. In recent years, industries of every kind have recognised this problem and are actively exploring and pursuing digital transformation, hoping to use digital technology to unlock the value of data and support long-term, sustained business growth.
So how do you get digital transformation right? The key lies in sound data governance, and it is not enough to emphasise technical support alone: bringing business value into play matters just as much. Only then can a good foundation be laid for the success of digital transformation.
I. The necessity of data governance
The building of data platforms in Chinese enterprises began roughly in the late 1990s, and nearly 20 years have passed since the first-generation architecture appeared. Yet the level of data development and platform building varies widely from industry to industry, and most enterprises overlook data governance. A steady stream of fundamental data problems, from inconsistent data to poor data quality, holds data platforms back and prevents data applications from delivering business results quickly.
Most enterprises face the following problems in data management:
1. Insufficient data standardisation: there are no unified data standards, which makes data hard to integrate and unify. The absence of a metric system leaves the relationships between metrics unclear; unclear metric definitions mean that the same metric maps to several different calibres; and master data is maintained independently in each business system with inconsistent rules and calibres.
2. Relatively poor data quality: the lack of quality management leaves huge volumes of data unusable because of poor quality. Without completeness, conformance and consistency, the conclusions finally reached are biased, and poor-quality data also raises related costs, both hidden costs and direct financial costs.
3. Weak data control: the absence of effective management mechanisms and unclear boundaries of responsibility between data management departments create barriers to cross-departmental data sharing and give rise to data silos, while long, drawn-out data acquisition processes make business analysis needs hard to satisfy quickly.
4. Simple data applications: the links between data are relatively weak, data has not been connected to the business system, and data cannot be explored or mined further, so its deeper value is hard to realise.
5. Lagging construction of data analysis platforms: some business activities lack informatisation support, a complete data and technology architecture and a centrally planned big data analysis platform, and the management processes of the entire big data platform are not managed effectively.
These categories of problems show that the issues facing traditional data platforms have not disappeared in the big data era; new ones keep emerging as well. Enterprises therefore need to further strengthen their data governance capabilities in order to solve these problems in the course of building big data platforms.
II. The goals of data governance
The goals of enterprise data governance are: to formulate data policies, safeguard data security, advance barrier-free data sharing within the organisation and ensure the smooth implementation of the data strategy; and to improve data management capabilities, optimise the organisation's level of refined management, raise the efficiency of business operations and strengthen the organisation's decision-making ability and core competitiveness, thereby providing solid support for achieving its strategic goals and, in turn, capturing data value, innovating business models and keeping operating risks under control.
III. Data governance solutions
1. Data standards management: each item of data involved in the course of enterprise management is given a standardised definition and a unified interpretation, together with unified definitions of the constraints and relationships between data, business rules and data quality requirements, so as to better support business development and system integration and to guarantee the consistency and accuracy of the data used and exchanged both inside and outside the enterprise.
2. Data quality management: data quality is supervised across the entire data lifecycle, and data quality problems are detected and resolved promptly from multiple dimensions, safeguarding the completeness, consistency, accuracy, timeliness and legality of data so as to raise the enterprise's overall data quality.
3. Data security management: a systematic data security control strategy is established, and comprehensive data security control mechanisms are put in place through user security management and data security management. Data security is delivered by combining technical measures with management measures, so that it can be managed in advance, controlled as it happens and audited afterwards.
4. Metadata management: the relationships and threads between metadata items are clarified, and the full-lifecycle processes of metadata design, implementation and operations are standardised, reducing the difficulty of using metadata and improving the user experience.
5. Master data management: the most essential and most shareable data is consolidated from multiple business systems and cleaned, enriched and distributed centrally, ensuring a single, accurate and authoritative source of master data.
