Big data dilemma for banks

Right now, digital transformation is in full swing. An optimistic result of digital transformation is to fully release the productivity of data elements. However, it is easier said than done, and it is not so easy to actually do it. Just imagine, which organization or organization do we have? Who dares to say that we will be able to achieve this goal?

Let’s go back to 2011. In that year, McKinsey released the report "Big Data: The Next Frontier of Innovation, Competition and Productivity", which ignited the wave of big data. Some leading banks took the lead in launching big data applications. , We have indeed seen the great progress that banks have made in data application. Data has played an important driving role in the digital transformation and digital operation of banks, and has exerted greater value. However, at the same time, the bank's big data application work is becoming more and more chaotic, giving people a feeling of seeing flowers in a fog.

The application of big data is a new job for banks, and banks are also moving forward in exploration. With the advancement of applications, the problems existing in the application of big data in banks are becoming more and more obvious.

First, the positioning of the data organization is not clear. Some banks that were early in the field of data mostly use independent data departments as the carrier of data applications. This model has indeed played a relatively good role in rapidly promoting data applications, and this model has also received more and more attention. Bank recognition, under the wave of digital transformation, many small and medium-sized banks have set up independent data departments. However, so far, the organizational positioning of the data department is not so clear even in some first-mover banks. There are two reasons. One is that some bank executives may not have enough understanding of the difference between data and traditional information technology, and they still position it as a traditional information technology department. On the other hand, for the banking business department, out of departmental interests, it is There will be a certain gap between the data department and the data department, which often leads to a sense of helplessness in the data department; the second is the confusion of the data department's own positioning. The managers of the bank's data department are basically from the information technology department. Even open recruitment tends to introduce managers with pure technical background. Because many managers and engineers with technical background have heavy thinking and lack of understanding of business, many data departments are building technology platforms in full swing. However, in terms of data applications It is basically driven by the needs of the business department. The technology platform is the basis of data application. This is understandable. However, the core of data work is the release of data value. Positioned.

Second, most data applications are still in the traditional analysis and mining stage. A lot of data application work in many banks still stays in writing analysis reports, doing traditional data analysis and mining work such as logistic regression and decision tree models, over-reliance on structured data, and insufficient unstructured data processing capabilities. For cutting-edge The overall application of artificial intelligence is relatively small, and there is also a lack of sufficient capacity reserves. Even if there are some applications, they are mainly concentrated in the field of customer service, and most of them are mainly imported from outside.

Third, the data department of the bank is not open enough. The data itself has connection properties, and the data application work must be open, inclusive, and integrated into the social ecology. The key is the openness and integration of ideas, and the ability to keenly predict the trends of society, industries, and technological development. Although banks have opened up a lot with the advancement of digital transformation in recent years, the data department, as an internal department serving the development of banking business, considers factors such as data security and commercial confidentiality, and its own openness is still far from the ideal state. In addition, the overall work of the head office of the bank is relatively stable, and some people are relatively closed-minded, lacking curiosity about new things and new technologies.

Fourth, the channel for realizing the value of data elements is not smooth. Data application needs to form an end-to-end closed-loop capability in order to effectively promote the application of data, the monitoring of results, and the iteration of models. However, the bank's organization chain and system chain are too long. On the one hand, the data department is at the back end, and it is difficult to directly touch On the other hand, it is inevitable that there will be one or another breakpoint in the long chain of circulation. The final result is that the data application is difficult to achieve the expected effect, causing the business department to question the data application work, and the data department is dissatisfied with the application enthusiasm of the business department, thus further aggravating the difficulty of data application.

Behind this problem is the lack of data strategy of the bank. From a strategic point of view, the business strategy is the core strategy of the bank, and the data serves the business strategy. However, there is still a lack of data strategy under the business strategy to guide the application of data. Work. Some friends may say that the bank already has a fintech strategy? Indeed, many banks have released fintech strategies, and data is included as part of the fintech strategy, which is also a manifestation of the lack of awareness of the bank's senior management on data work.

As a new type of production factor, data has its unique characteristics compared with traditional information technology, which are mainly manifested in four aspects:

One is the strong business attribute of data. We have mentioned the business attributes of data in the previous article. Data provides solutions to business problems that cannot be solved by existing means. It is an active business innovation activity. Traditional software development is to use software to determine business solutions. Realization is a relatively passive execution operation. Therefore, the core of data application work is innovation, which leads to business;

The second is the strong connection property of the data. Data in the database is a field, and the data can only be maximized if it is connected as much as possible. Just like if we only look at a field like "loan amount", the information that can be obtained is limited and the value is relatively low. If we By extending the data fields to the entire database table, the information and value we can obtain will be much richer. Similarly, if we connect the bank data with the data of the Internet, the government, etc., the value obtained will be much greater ;

The third is the strong research attribute of the data. We say that data is to find answers to problems that cannot be solved by existing means. This depends on several situations. Some problems can be learned from existing solutions in the industry, and some problems are not readily available in the industry. The solutions can be used for reference. We need to explore and explore by ourselves. Theory guides practice. In the process of exploration, we often need to find directions from theory, and this theory is not entirely at the algorithm level. Sometimes you may You also need to find answers from the theories of finance, sociology, psychology, etc. Sometimes you have been busy for more than half a year, but you may end up with a failure. Therefore, data work has a natural research attribute, and some people call it for "Data Science";

The fourth is the strong refresh property of the data. This is mainly divided into two levels. One is the data level. The fresher the data, the faster the update, the more value it can generate. The so-called freshness means that the data is relatively close to the current practice and the frequency is high. Data is the mapping of the real world in the digital world. The real world is changing dynamically. As time goes by, people's behavior habits and social operating logic will change. As humans enter the digital age, this change will become more and more serious. Faster and faster, so you can't use user portraits made with data 10 years ago in today's marketing. The other is the algorithm level. Today’s algorithms are updated very quickly. Only by keeping up with the beat and refreshing yourself can you keep up with the times.

Data represents new productivity. We all recognize the important mission of data in the future. We must attach great importance to the unique characteristics of data work and establish independent strategies that adapt to these characteristics to guide the direction of data application, form data ideas, and promote the establishment of The corresponding management mechanism and working methods can no longer put old wine in new bottles, and still use traditional information technology management methods to promote data application work. As far as the current situation is concerned, although the bank's current data application has achieved certain results, after several years of continuous traditional data analysis and mining, the marginal utility of the pioneers has dropped significantly. , for banks lacking a data strategy, the marginal utility of data applications will quickly approach zero, and it will be difficult to achieve a qualitative leap. In the end, they will fall into the trap of "stagflation" and may even become tasteless. This is by no means alarmist. When you From 0 to 1, you can do whatever you want to generate value. When you need to go from 1 to 10, what you need is deep precipitation, deep thinking, calm prediction, and orderly advancement. Without a strategy, you will never get out of chaos. Therefore, for leading banks, it is time to rethink their data work, and for banks that are just starting out, it is time to establish an independent data strategy from the beginning. As for how to build it, we will discuss it below.