Translated by
2020/01/13 17:21:45

Svetlana Bova, Chief Data Officer of VTB bank - about the main issues and errors in data management

The relevance of a subject of management data (Data Governance) grows every year. Really, need of the organization of the processes directed to increase in efficiency of collecting, processing, storages and use of data as a valuable asset is already obvious practically to all companies. It is much told about what advantages bring to the company correctly built processes of management of data, and many organizations already began implementation of this initiative. At the same time the organizations often make similar mistakes which have negative effect on rates of implementation and efficiency of the created processes of management of data. About what it is errors as to avoid them and on what questions the organization should find answers in the course of Data Governance implementation, in the material prepared for TAdviser tells Svetlana Bova, Chief Data Officer bank VTB.

Content

Why companies data management?

Svetlana Bova: it is impossible to try to improve quality of all data in the organization at once at all
Svetlana Bova: it is impossible to try to improve quality of all data in the organization at once at all


Sometimes the organization begins implementation of processes of management of data, without understanding ultimate goals and measurable indicators of their achievement. Also there is no assessment of the current level of a maturity of the organization regarding data management and the strategy of development of function.

The initiative can be dictated by external factors, such as general trend on digital transformation, implementation of processes of management of data in the companies competitors ("at them is, too it is necessary for us"), at the same time internal premises are absent or are opaque to company management. Thus, there is an attempt of implementation of what the company is not ready to. Moreover, perhaps the company also does not need it at the moment. Naturally, perspectives become quite foggy.

How to avoid it?

It is necessary to put accurate and measurable business - the purposes. For example, to increase efficiency of the marketing companies by 10% due to quality improvement of client data, to reduce by 30% costs on implementation of IT systems at the expense of the description of architecture of data of the organization, to lower by 50% of a labor cost for verification of financial and risk statements due to standardization of business terms and creation of the centralized data mart, etc. These purposes should be clear, confirmed with business divisions and company management and are reflected in the strategy of data management.

What to begin with?

It should be noted that it is impossible to try to improve quality of all data in the organization at once at all or to describe all terms used in the company in the business glossary. Concerning data management, it is possible to formulate the Pareto principle as follows: "20% of all data critically influence 80% of business processes of the company".

The following sequence of steps will be the correct approach in this case:

It is necessary to reveal the most critical processes in terms of business, to define the data significantly affecting efficiency of these processes, to record initial efficiency evaluation of processes. Further consistently step by step to raise the quality level, data availability and the description of their metadata. Upon completion of works to estimate as far as it was succeeded to reach the stated business objectives and to pass to activities on the following business process, scaling thus activities for system work with data.

What the main thing in data management?

There is a certain stereotype that implementation of expensive IT systems is enough to bring order to data. So if to implement, for example, the industrial data warehouse, Big Data or management tools data, then with data everything will become good at once. By my experience, it not so. Practically everyone who dealt with quality of data, knows about the principle of GIGO (garbage in - garbage out). This principle speaks about if malformed data come to point of entry in the IT system, then at the exit the result will be useless even if in the IT system all algorithms fulfilled correctly.

Therefore I select the priority directions of works on data management following as the importance:

  • People – the most important resource. Employees should have a culture of work with data, understanding that data are a valuable asset using which the company can get an additional profit or reduce costs, and desire to be owners of data.
  • The processes regulating collecting, processing, storage, use of data, works on quality improvement of data and maintaining the glossary of business terms of the company, architectural standards, institute of ownership of data and so on.
  • Technologies: the business glossary, tools of data flow control (Data Lineage), instruments of integration, design of data models, the software on quality management of data, management systems for master data (MDM - Master Data Management) and management systems for the normative reference information (RDM - Referential Data Management).

Of course, technologies are necessary, but nevertheless they are secondary in relation to people and the organization of processes.

Who such owner of data?

Happens that management of the organization considers data management a technology initiative. Data are stored in IT systems, it means, as IT specialists should be responsible for them (to own them). This quite misleading statement. The concept when the divisions using data in the processes, for example, for preparation of financial statements are appointed owners of data is also often applied. So the one who is most interested in their quality owns data. Such approach is possible, but nevertheless it is not rather effective.

If, for example, the consumer buys low-quality goods to whom he will make a complaint in shop? Of course to the producer of goods. And with data: them only the one who creates them, in what process data are born on light, the one who can directly affect efficiency of a production process of data can own. Including to regulate processes of data entry in the IT systems, processes of quality control of data as automated, and manual, to initiate necessary completions of IT systems.

What should be CDO?

CDO is considered a heavy profile of candidates for search and hiring, and to it there are several reasons: first, in Russia practice of selection of activities for data management in a separate functional unit of the organization exists not so long ago, secondly, the educational institute by this profession in difference from the international community is not built yet, thirdly, statistics of successful people on this position for this reason so ambiguously and difficult happens to make requirements to the candidate for a position of the leader in data management in each specific organization is not saved up.

Sometimes, the company defines for itself that technical competences are the main for the Chief Data Officer. Yes, technical competences are necessary including knowledge in the field of data warehouses, instruments of data translation, skills of design and so on. But, world practices prove that another is much more important. CDO is first of all the strategist who determines the strategy of the company by work with data. CDO is a businessman who creatively offers business of a way of increase in process performance at the expense of data management. And eventually, CDO is a missionary whose task to turn everyone into the organizations in the belief: belief of importance of data as valuable asset.

Therefore the main necessary CDO competences can be provided as follows as the importance:

  • The high level of communicative skills, including emotional intelligence, charisma;
  • Strategic thinking, ability to implement innovations;
  • Knowledge of data domain and business processes of the company;
  • Technical competences.

When does the initiative of data management come to an end?

There is an opinion that data management is the final project, single exercise on cleaning and inventory of data. However it not so. The world around promptly changes, data disappear and appear, information systems migrate, requirements to data change. Therefore after implementation of a framework by Data Governance works on data management should be continued in the regular daily mode within operating activities of the company. Continuous work on improvement of collecting, processing, data storage, increase in their quality and penetration of the idea of the responsible relation to them becomes an integral part of life and a company culture. Only then it is possible to speak about success of implementation of an initiative of data management safely.