Tool, automatic scoring and entry into the Asian market: what trends dominate in online lending

[Carding Forum]

We looked at how the introduction of advanced technologies helps financial companies grow.

Despite the coronavirus crisis, non-bank financial institutions were able to increase the volume of transactions in 2020. So, according to the NBU, almost all market segments showed significant, and in some places - record profitability indicators. However, there is no need to rest on our laurels: in the context of widespread digitalization without the use of online products, they will be squeezed out of the market. Players in all countries understand this. For example, in December 2019, Halyk Bank of Kazakhstan launched online lending to meet consumer demands. The innovation turned out to be in demand.

“In 2020, the share of online loans through a mobile application in terms of quantity was 34.2%, and already in 2021 - 57.6%,” Dinara Amanbaeva, Director of the Development Department, said during the international conference “Digital Credit Technologies” direct sales and premium services to Halyk Bank of Kazakhstan.

At first, authorization in the bank's application was available only to the clients of the financial institution, now it is open to everyone. According to Amanbaeva, the bank is constantly working to simplify the lending procedure on the Internet.

“For example, a user chooses an online loan or refinancing,” she continued. - Further conditions are offered to him. After reviewing them, he clicks "get a loan" and independently chooses the amount, term. Seeing the monthly payment, he calculates his options. If the amount and term is suitable for the client, he clicks the "Next" button. Then we conduct an internal scoring of the client. If it does not suit him, then we offer additional options. Online credit is automatically credited to his card account."

Installer loan rate​

In addition to the convenience of the procedure, the financial company must offer the consumer the most optimal loan product. Olesya Kiseleva, managing director of Lime Credit Group, a company that specializes in short- and long-term online loans, advises to bet on installment loans. These are loan products that allow the borrower to repay the amount gradually over a certain period of time. In Lime Credit Group, more than 60% of the portfolio is made up of just such loans.

“The client can take the same amount both with the condition of one compulsory payment, like an ordinary short-term loan, or pay in equal installments,” says Olesya Kiseleva. - By offering a client with a high score by installments, we give the opportunity not to multiply their debt load within one month, but to pay off within three months. At the same time, we leave him the opportunity to repay the loan ahead of schedule."

According to her, the profitability of such loans is higher. In addition, the company has developed an approach to optimizing the cost of providing loans.

“At the stage when we buy traffic - to webmasters, to lead generators, who are the main source of clients for most MFOs, on the basis of sorrow we estimate how much we can offer. Based on this, we diversify the cost of a lead. The better the client, the more we can pay for it, thereby we set up our partners-lead generators to send us only the highest quality clients. We understand that having paid such a cost, we are more likely to earn more, "said Kiseleva.

At the same time, she continued, if the basic checks give reason to make a positive decision to issue a loan, the firm does not spend money on bypass verification.

Scoring - on the machine​

To make a decision on the issue of money, lenders, as a rule, use data from the application, credit bureaus and anti-fraud databases. Less often they resort to alternative sources - data from mobile operators, social networks. And there are very few of those who use behavioral patterns (templates) when filling out an application or biometric data (voice). Although information can even be provided by an email address, noted Anton Astrakhantsev , Chif Risk Officer Gofingo.

“As the Regional Head of Fraud Prevention, we created over 50 variables based on just one field - email address. It should be recognized that the informative value and weight of the trait could vary depending on the country. However, it can be strengthened by cross-checking with other sources of information. At the same time, some of the found variables were quite predictive regardless of the country, ”he said.

It is a thing of the past to assess the borrower's creditworthiness (scoring) only in manual mode. Alena Andronikova , CEO of Moneyveo, says: in 97% of cases, the decision to issue a loan in a company is made by a machine, only one loan at the same time undergoes two hundred automated procedures. The photo-verification process is also built on the basis of artificial intelligence (AI).

“When we send a client to an operator (in 3% of cases), we collect copies of documents. The system recognizes the type of document, and for example, it can compare a selfie with a copy of the passport available at the credit bureau, ”said Andronikova.

Eastern vector of development​

Having reached a certain level of development, financial companies are thinking about expanding the boundaries. According to Roman Katerinchik, founder of MyCredit, given the scale of the markets, it is better to choose Asia for expansion. Having received investments from a banking group in 2019, the Ukrainian firm decided to enter the Vietnamese market. It was possible to gain a foothold there, but it was not easy.

“Our mistake was that we followed the country management model. That is, a person from Ukraine runs an office in Asia and there is a local CEO. As a result, we got a large number of conflicts, - said Katerinchik. - You should take into account that the mentality of the Vietnamese has a super-clear hierarchy: the boss is in charge, and all tasks for the team must come through one person. Although we, as business owners, are used to communicating directly with the head of marketing, CEO. In Vietnam, it became clear: the most effective signal is spread when we communicate through one person."

Therefore, the work was structured in such a way that the decision is made by the local local management, and the headquarters ensures the smooth operation of IT and marketing. In Vietnam, verification takes place, there is a call center, technical support, copywriters.

Another mistake, Roman recalls, was to use Ukrainian KPIs. It turned out that they do not work there. Instead, competition is the best motivation.

“As soon as they drew a table of students' progress on the blackboard and began to swap them, depending on the results of the previous day, the team literally came to life. In addition, we realized that praise and love work better than criticism and reproaches, ”said the founder of the company.

Also, he continued, one should bear in mind the peculiarities of building a risk model in Asia.

“We use machine learning, collect a large number of data sources - databases, credit bureaus, contact mobile wallet operators, accumulate this data both in the anti-fraud process and in the verification process. This is a big challenge. After all, while banks have access to credit bureaus, it is more and more difficult for financial companies, ”said Katerinchik.

Therefore, in his opinion, it is better to build a risk model by actively issuing small amounts of loans to all new clients. Thus, it can be developed relatively quickly - in 2 - 3 months.