A perfect legal system can escort the booming development of the credit collection industry. China's current legal system for the credit collection industry mainly includes the Regulations on the Administration of the Credit Collection Industry and Measures for the Administration of Credit Collection Agencies issued by the State Council in 2013, as well as the Guidelines for the Supervision of Credit Collection Agencies issued by the Central Bank in 2015 in order to further clarify the operational norms of credit collection agencies. Compared with the mature markets in Europe and the United States, China's credit collection industry legislation is not sound enough and is in the early exploration stage, with only administrative regulations or departmental rules, both of which have lower legal effect. Moreover, China has not yet established a sound legal system for data collection and personal privacy, which has led to privacy protection problems in the collection of personalized data in the credit bureau industry.

Secondly, the number of people to be covered has to be increased;

Society's demand for credit information is huge, while the credit collection system is incomplete, and the government-led credit collection system can hardly fully meet the needs of the financial market operation. The coverage rate of credit information of enterprises and individuals in the United States is as high as 80%. Even so, credit bureaus in the United States continue to invest in the development of exclusive data sources, and enhance the depth, breadth and quality of the database through the analysis of new data, which provides a solid foundation for credit collection. And according to China's central bank credit system, as of the end of 2015, the personal credit system included 880 million natural persons, of which 380 million had credit records, 500 million had only simple identity information, and another 500 million or so were not in the central bank credit system. In fact, the target customer groups of consumer finance companies are mainly concentrated in the middle and lower class consumer groups, which are mainly young people, such as office workers who have just joined the workforce, and groups with low income, etc. The target customer groups of consumer finance companies are mainly young people, such as office workers who have just joined the workforce, and groups with low income. This part of the real need for consumer financial services user groups, precisely in China's central bank credit system in the lack of personal credit records.

Thirdly, the market penetration rate needs to be improved;

Data from Avery Consulting shows that the market penetration rate of China's personal credit industry is generally maintained at around 9%, and in 2015, the potential market size of China's personal credit industry was RMB 162.36 billion, while the actual market size was only RMB 15.14 billion. With the change of personal consumption and transaction habits, the application scenarios of credit are increasing. In addition to credit and credit card consumption, non-financial areas such as rentals, car rentals, shopping, visas, etc. also demand personal credit information, and the market penetration rate needs to be further increased.

Fourth, data collection standards need to be unified;

Data collection is the foundation of credit collection, for this reason, the U.S. National Association of Credit Management has formulated a standard data reporting format and a standard data collection format to standardize credit data and facilitate the enjoyment of credit data among agencies***. However, the lack of an effective *** enjoyment mechanism for various types of data in China has led to a serious problem of data silos, and the existing data is highly homogenized, mostly data available through public channels, with no personalized and exclusive data sources. At the same time, various types of data are uneven, the lack of uniform standards, directly affecting the quality of the credit report.

Fifth, the ability to analyze data needs to be improved.

Data analysis ability directly determines the quality of credit services, therefore, data analysis is a key link for credit enterprises to transform credit data into credit products. The data analysis technology in the United States started very early, as early as 1956, the FICO scoring system was launched, after more than half a century of continuous improvement, the application has been very extensive. At present, more than 90% of large credit bureaus, including Experian, Equifax and TransUnion, have adopted the FICO scoring system. 2009, the U.S. company ZestFinance took the people whose credit scores were too low or whose lack of credit history resulted in abnormally high borrowing costs (FICO scores of 500 or less) as the target of its service, and in its credit evaluation analysis, it integrated multi-source data and introduced machines. It integrates multi-source data, introduces machine learning predictive models and integrated learning strategies for big data mining.ZestFinance's core competitiveness lies in its data mining ability and model development capability. It is understood that 3,500 data items are often used in its models, from which 70,000 variables are extracted, and 10 predictive analytics models, such as the fraud model, the identity verification model, the prepayment ability model, the repayment ability model, the repayment willingness model, and the stability model, are used to conduct in-depth learning and obtain the final consumer credit score. On average, a new version of each model is created in six months to replace the old one. The new version usually includes more variables and data sources. zestFinance uses algorithms from Google's Big Data model. Thousands of raw data from third parties (e.g., phone bills, etc.) and borrowers are entered into the system to find correlations and transform the data, recombine the variables into larger measures based on the correlations, and finally input these larger variables into different data analysis models, where the conclusions from the outputs of each model are formed into a final credit Score. Compared with traditional credit management business, ZestFinance's processing efficiency has increased by nearly 90%, and in terms of risk control, ZestFinance's model has improved its performance by 40% compared with traditional credit assessment models. On the contrary, in the domestic credit collection industry, data analysis has just started, and the efficiency and accuracy of data analysis need to be further improved