Portfolio /
Automation of financial organizations
‘USSD-banking’ for microfinance company ‘SMSFinance’.

USSD – is a technology that allows you to organize high-speed exchange of information between the subscriber’s cell phone and the service application, in real time.

The project aims are to relieve operators by organizing direct communication of customers with the service, through the USSD Protocol, and increase the conversion of loans through the connection of a new alternative method of processing loans.

 

As a result of the project, the clients of IFC ‘SMSFinance’, whose personal data was already in the system, were able to access the services of the credit institution from a cell phone without the participation of the operator. The system allows you to issue a loan and get it through one of the channels-QIWI, Contact, Leader, Yandex-money using a credit card or Bank transfer. In addition, you can repay and extend the loan through the QIWI system, request background information on the status of the debt or leave a request for a call. The service has become even more affordable and convenient!

Personal account

Personal account (PA) is a key component in the ecosystem of the credit organization SMSFinance’. Every day, customers from all over Russia use it to find out their debt on the loan, make another payment or apply for a new loan.

Technically, PA is a web application that can be accessed from any modern desktop and mobile browser. The old version was written more than 5 years ago and until now has not changed dramatically. Over time, it became obsolete and began to lose ground in comparison with similar solutions of competitors.

 

At the beginning of 2019, the management of Smsfinance set a task to release a new Personal Account of the client.

The criteria for success identified the following:

  • 1. Friendly user interface. The system must understand what the client is currently waiting for from it and offer him the next step to achieve the goal.
  • 2. Responsive design. The site should be adapted to mobile devices-phones and tablets.
  • 3. Support for new products in PA.
  • 4. Moving from legacy tools to a new modern technology stack. Ensuring system operation without reloading pages.

Before improving the interface, we decided to prepare the ground for functional changes. Unlike life, the statement “the old friend is better than the new two” is not applicable in terms of technology. HippoCMS, Wicket and jQuery tools have been eliminated in favor of modern Angular 7, TypeScript. We replaced the outdated session authorization with authorization through OAuth 2 tokens. Thus, we began with point #4.

Laying a good foundation for the development, we gradually built the functionality: worked UI/UX of the main pages of the personal account, for all pages of the LPA implemented adaptive design: all our styles applied to the components are adaptive to any screen size, including mobile phones and tablets.

For the convenience of clients, we have built a seamless process of issuing a loan: after registering on the site, the client enters the personal account without entering a login and password, the system prompts the client for further steps to obtain a loan:

  • View available credit offers and current promotions.
  • Choose the appropriate loan product, amount, loan term and method of receipt.
  • Read the individual conditions.
  • Sign a loan agreement.
  • Get money using the chosen method.

After receiving the loan, personal account becomes a real assistant for a client: displays information about the debt, informs detaily, what the amount of debt is, shows the schedule of planned payments and the documents on the loan, suggests, in what terms it will be cheaper to repay the loan to the customer and how to keep good credit history if the client cannot repay the loan on time.

Thus, we confidently fulfilled points #1 and #2.

As for paragraph #3, in 2019, changes in the work of microfinance companies came into effect. Now the maximum loan rate should be no more than 1% per day. In order not to face losses, MFIs need to change their business model, reorienting it to long-term loans. “The” old ” personal account was not adapted to multi-product offers: the client could only take a short-term loan. Adding new products was time-consuming, requiring the allocation of a large number of developer resources. The new PA supports the rapid addition of new product offerings, with minimal developer involvement. The customer has a choice of several product offerings.

Seamless registration

The classic registration process is as follows with:

  • The user registers in the system
  • The system sends SMS with the password from the personal account (LC).
  • The user is authorized using the received password.

When registering, a significant portion of users are lost in the authorization step. This reduces the conversion rate for registered users. Business gets fewer customers, which means revenue falls.

 

We have implemented a seamless registration solution for a credit institution. After the implementation of the solution, the process was simplified. SMS with password is no longer sent, re-authorization is not required. When the user fills in the registration data, scoring starts-an automatic process of checking the reliability of the borrower. By the time of completion of the questionnaire, the scoring result and the credit decision are ready. If this decision is positive, the user has to choose the method of obtaining a loan, after which the system instantly transfers the money to the client’s account.

Seamless registration is part of the implementation of the onboarding strategy, i.e. retaining the client and bringing him to the successful implementation of the business scenario laid down in the system. Customer satisfaction has increased, conversions have increased, the business has received more revenue.

The following technologies were used:

  • Technology frontend: Angular 6 .x, Webpack, nodejs , sass, lodash, Nginx
  • Server technology: Spring Boot 2.x (spring mvc , spring core, spring data jpa, spring security), Tomcat 8.x
  • Database Postgres 9.x
The customer identification

As part of the development of the public services portal, a unified identification and authentication system (ESIA) was created in the Russian Federation. It is a Federal system designed to provide authorized access to information. ESIA is implemented with the help of solutions certified by the FSB, which guarantees compliance with the requirements of legislation on the protection of personal data. Later, by order of the government, ESIA was integrated with Federal portals, websites of the Executive power of the subjects of the Russian Federation and other resources to provide priority services. The same government order allowed private organizations to integrate their information systems with the ESIA.

When providing electronic services, it is important to ensure information security. We have integrated ESIA with the personal account (PA) of the borrower of the financial company. As a result, two indicators increased. The conversion of the registered users and the conversion rate on loans. The first became possible, because the entrance of the ESIA increases the confidence of users, the second-due to the availability of more channels for issuing loans in the identification of the ESIA. Work with the service has become more convenient. For users who already have an account in ESIA, now you do not need to register in the PA – only login. In addition, you do not need to enter passport data and fill in other fields that are stored in the ESIA. The system automatically transfers information from ESIA to the system.

 

The project used technology:

  • Technology frontend: Angular 6.x, Nginx
  • Server technologies: JWT/OAuth2, Шифрование RSA SHA 256, Spring Boot 2. X, Tomcat 8. X
Risk Engine – a solution to optimize the scoring and maximize the profit of the credit institution (CI)

Fast and accurate assessment of the solvency of borrowers (scoring) is a key task of the CI, on which the profitability of the business depends. There are many scoring services on the market. Different scoring services perform analysis of borrowers ‘ solvency according to different parameters, thus, the results of inspections complement each other. As a rule, to decide on a loan, KO consistently refers to several scoring services. The quality of scoring depends on the correctly selected scoring services and the correctly set sequence of checks, which means the profitability of the CI.

 

It is impossible to maximize the financial efficiency of scoring at once. The quality of scoring is improved iteratively by analyzing the results and making adjustments. Changing this process is a time-consuming and expensive operation that involves programming, testing, and debugging. For these reasons, in many CI, this process is not optimized and works with low efficiency.

“Risk Engine” is a tool that allows you to quickly change the scoring process.

The results of the implementation of “Risk Engine” in one of the largest IFC in Russia – “SMSFinance” are presented in the table below.

Before Risk Engine – After Benefit
Connection of a new scoring service took from several days to several weeks. Connection of the scoring service takes several minutes. Increasing the speed of decision making.
Reduced costs of change.
Problems in one scoring service lead to incorrect scoring results. Problems in one scoring service do not affect the scoring result. Scoring process is failure-free and correct.
Analyzing financial efficiency of scoring services is labor intensive. Analyzing financial efficiency of scoring services is easy and transparent. Reporting on each scoring service separately is enabled. Enhanced efficiency of the scoring process.
Reduced costs of scoring services.
Simultaneous support of several scoring processes is labor intensive and generally not used. Now it is possible to quickly launch several scoring processes simultaneously and compare their actual profitability. Main process handled by majority of lenders. Experimental processes handled by a limited number of clients. Scaling tested efficient experimental scoring processes.