Not long ago, Hingkat published an article on consumer finance (Portal: "These three questions may be the biggest myth of the consumer finance industry.") Whether the scene of the consumer finance industry is valid for wind control , The downturn in the business of commercial banks and rising operational costs of consumer finance business, put forward some views.
After the article was released, there are many readers to communicate with me in the background, especially the first one, which are the strengths and weaknesses of the online and offline scenarios of consumer finance, Point of view, gave me a lot of practical experience.
Recently, I also took these issues and talked a lot with different types of consumer financial service providers and practitioners:
✔ The existence of the scene is of course valuable, but "Scene King" is really a myth. A company with a scene does have a role in getting customers in bulk, reducing acquisition costs and controlling the flow of funds, but business-to-business puts higher demands on the company's ability to manage the scene. Therefore, for wind control, the more fundamental logic is still the choice of customer base, the building of capabilities rather than the scene itself.
✔ Returning to the construction of consumer finance company's risk control capability, the core is to really optimize and rebuild the traditional process through the application of new technologies. Many companies have introduced technologies such as smart customer service, face recognition and smart collection to improve efficiency and accuracy. However, it is still a problem for many companies to determine whether these capabilities are introduced into third-party or independent research and development.
✔ However, consumer financial companies for the layout of the scene, in fact, there are other considerations outside the risk control and access. Because of different scenes, product design and data accumulation of different people are all different. In the meantime, the layout of different kinds of different scenes plays a positive role in collecting data, enhancing the product experience and even the financial performance.
1
The view that most practitioners agree is that in the consumer finance business, the customer base is the most important factor. By analyzing the customer base structure of each customer, the basic risk status can be judged and the scene can only be used as a One of the judgments.
Just like the product of Ant Financial, "Micropay Loan" is a non-scene-based online lending product, even a bank credit card user and a bank's consumer credit whitelist users are all There is no way to control the flow of funds, but because the user base is of sufficient quality, the bad debt ratio can be kept at a very low level.
On the contrary, large-scale risk events have erupted in such fields as education, medical and beauty. The problem is not so much the scene itself. It is too dangerous to rely too much on the scene to ignore the judgment on the "human" itself.
In fact, if the ability to control wind control is too dependent on the scenario, it is equal to the flow of funds being controlled by others. This is an unconfident measure of the company's ability and a detriment to the long-term development of the company.
At this point, I quite agree with the immediate consumer financial CEO Zhao Guoqing's point of view, " a truly capable company should not rely on the scene of risk control, because we do not give credit to the scene, we give credit, credit is what is It is dependent on the data behind him to make him trusted. "
From this perspective, the significance of the scenario to risk control is more important than the control of the flow of funds, which is actually the acquisition of customer segments.
Whether it is online or offline, high-quality guest scenarios and traffic are conducive to the company to obtain higher quality users at lower cost, thus occupy more effective data to enhance the level of wind control and maintain a low level of bad debt, these advantages are conducive to Platform access to lower-cost funds favor and promote profitability continue to improve, while further reducing interest rates will help enhance the user's sticky, into a virtuous circle.
Of course, this is also a "virtuous circle" in a relatively ideal situation. Many companies are facing the reality that except a few e-commerce giants dominate the online trading scene, some consumer finance companies can rely on shareholder resources to monopolize offline offline scenarios , Most consumer finance providers still mainly rely on embedding the scenario by cooperating with third parties, which means that the quality of the customer base facing most platforms may be similar.
And if the scene, the crowd is highly coincidental, the scene of the consumer finance battle may only determine the "amount", and the real decision "quality" or depends on who can long lending crowd, capital flow control more in place, more sophisticated, This is in itself a matter of ability.
2
Many people still think this is a "retreat" topic. Everyone is talking about big data and artificial intelligence. How exactly they are applied and how they look like they have a "siege" barrier With people outside it is always difficult to understand.
In fact, after I spoke with many companies, I felt that these technologies were applied at different stages of the business process, and many of the optimizations could be just a small step that the user might not even feel at all, but the data The improvement is remarkable.
Cite a few examples, there may be more intuitive experience:
For example, when they came in contact with the AI team of consumer finance companies immediately, they had been independently developing face-to-face systems with an accuracy rate of 99.99%. According to the consumer financial measurement results immediately, in practical applications, the product may be higher than the domestic face recognition of other products a few percentage points higher.
In fact, it is hard for ordinary people to feel the difference between the passage rates of one and two percentage points. However, it is of great significance to these consumer finance companies. It is understood that the annual trading scale of head enterprises in licensed consumer finance companies can reach tens of billions of yuan. If the fraud risk can be reduced by one percentage point, it will reduce the loss of several hundred million for the platform.
Similarly, artificial intelligence customer service for the cost savings, but also this truth. Most of the companies that previously provided consumer financial services had a large number of call centers, which were responsible for a series of work such as credit audit, customer service and even collection, which could reach thousands in a large scale. However, right now, smart customer service has begun to replace artificial customer service.
Prior to the data released by Tencent, Micronesia's smart customer service and customer service 8 artificial, the daily message processing capacity reached 1.8 million, which is equivalent to 800 daily customer service for 10 hours of telephone connection workload, the ant gold clothing More than 90% of customer service work is done by artificial intelligence.
At the same time, the accuracy of smart customer service is also getting higher and higher. Zhao Guoqing said that immediately consumer finance proprietary proprietary natural language parsing engine intelligent customer service system, the number of calls back a day also exceeded 100,000, with the ability of e-commerce smart customer service, the accuracy rate of more than 90%.
What is worth emphasizing here is that many consumer finance companies' technical services are provided by third parties, as embedding proven technologies into the scene can quickly bring new technologies to the application stage, serving the products and users. However, a new trend is that many platforms with user and data advantages have begun to explore proprietary R & D technology systems.
Still face recognition technology, for example, for the development of artificial intelligence, data, scenarios, science and technology are indispensable for all three. Especially when the sample data is large enough, not only the accumulation of positive samples, but also the feedback of "bad seeds" is needed. However, through the embedded services provided by third-party service providers, the negative feedback can not directly affect the learning engine , It may be difficult to improve the accuracy even further.
"On the AI, some data companies lack the technology, do not know how to use the data; there are some technology companies lack the scene." Immediate consumer financial CTO Jiang Ning said that "immediately consume the money is both, So we hope to do a better job in this area through independent research and development.
3
The discussions on the scenarios of consumer finance, wind control, and capacity building are in fact a matter that is constantly changing in the course of development. It is hard to draw a conclusion. In particular, the development of the industry has come to a new stage, some companies will also have some new considerations in the layout.
Before many companies believe that online and offline products and targeted at a completely different crowd, the wind control logic is very different, it is difficult to achieve full coverage.
However, there are also many companies that have extended their online and offline lending to improve the user experience after they have accumulated enough data and users in the early stages, or expanded their user base to enrich the online user experience. Data and products, and even further play the role of financial defense.
However, the more complex the product, the more diverse the customer base, the higher the demand for the ability of the consumer financial company. Especially after the traffic bonus period, these companies still have a long way to go.