Case koritsa: how to make money on the evaluation of foreign customers

“We just decided to make with the kids”

In 2013, the graduate of the Perm Polytechnic University and MBA at Istanbul Yeditepe Üniversitesi, Weichman Mary turned at once three Russian companies, which asked her to write a business project to open microfinance institutions (MFIs) and to develop a scoring system. “I decided that the demand is there — there were no such services for MFIs, the organizations themselves were not connected to the credit Bureau, few people knew what a risk assessment”, — says Maria.

Scoring in General form looks like this. The system receives information on the borrower profile, credit history, data from the different databases of state agencies, etc. the system Then processes this information and generates assessment score. If he is passing — the decision on issuance of credit.

Three prepaid order (each about 600 thousand rubles), made it possible to gather a team of four mathematicians and friends to start developing. “In fact, it was such a cabal, we just decided to earn money with the boys” — laughing, Weigman. In February 2014, “Skorista” drew the first €100 million of investment from a business angel, a former Bank employee. MFO then relied mainly on logic and experience, and we give them a suggested mathematical model,” says Weichman.

The market MFIs grew rapidly. According to “Expert RA”, the aggregate portfolio of loans to MFIs amounted in 2013 to 39 billion roubles and in 2015-m 63 billion. the Total number of organizations at the end of last year, according to the Central Bank, amounted to 3688. For the most part this is a small business that cannot afford to develop their own scoring models. “Skorista” suggested MFIs to give this function to outsource is to evaluate borrowers for receipt of applications.

Investors and accelerators

During the spring and summer of 2014 there was an active growth “Christy”. The startup has processed the applications of MFIs and increasing client base. The performance impressed the venture Fund Life Sreda invested in June 2014 in the project $400 thousand In that time the project team had less than ten people, some of whom worked in the office in Perm (programmers and rankovici), and a part — in Moscow.

The problems began in the fall of 2014, when Russia began to gain strength the financial crisis. “The algorithm just stopped working, complains, Weigman. — People stopped paying on loans, it became clear that the need to rebuild the entire system in principle”.

Team entrepreneur have gone a mathematician, and behind him and the programmer, which frightened the outbreak of difficulties with financing. “They understood that we are finishing the last of his money,” recalls Wachman. — And confirm the efficiency of the algorithm no.

The founder of the &lsquo;Christy&rdquo; Maria Weichman<br />
&nbsp;
Photo: Oleg Yakovlev/RBC

“>

The Founder Of “Christy” Maria Weichman
Photo: Oleg Yakovlev/RBC

Saved the company another round of investment from Life Sreda in the amount of RUB 12 million, which the company received in February 2015. Then the team was joined by a new risk Director, with whom, Wachman in June 2015 made a new algorithm for evaluating borrowers. “We used the same sample, but other variables describing the borrower more accurately. Variables are created from various data, it is not only credit history, profile of the borrower, but also data from public sources, payments for the phone. Simple variables are calculated and aggregated. For example, the ratio of the sum of closed micro-credits to the amount of the loan,” says Weichman. According to her, if initially the system took into account 500 variables, but now their number reaches 7000.

By mid-summer of 2015, the company has monthly treated about 45-50 thousand applications from MFIs, which allowed to reach self-sufficiency.

Autumn “Skorista” hit the acceleration program FUND. “We made a working algorithm, but have difficulties with sales, recalls, Weigman. — IIDF, helped us to find a new business model. We have organized support that was better to sell than the people responsible for the sale”.

According to the register, 43% of shares of OOO “Skorista” belongs to Wahman, 40% — OOO Smartmarket” (100% Vladislav Solodkiy, managing partner Life Sreda), 11% from Alexey Pure, 4% Ivan Tretyakov and 2% in FIDI.

Scoring economy

MFIs tend to create their own systems of risk assessment. “The model created in the private data of the company more efficient than those that are created on the data market — said the Director of risk MFIs “Migkredit” Artem Bykov. — We independently processed tens of thousands of applications that use multiple scoring models to evaluate various customer segments. This is behavioral, sociodemographic, and fraud model.

But to afford its own scoring model not many people can afford. Such MFIs and work with “Christou”. Sales are mainly carried out directly: the company’s employees call up MFIs in the registry and offer to try a scoring model. Well-proven industry events and word of mouth — customers often exchange information with each other. Today “Christy” 60 regular customers, among which the “Second money”, “Blitz”, “Honestly”, “Karmana”, “Optimine,” etc., says Weichman.

Outsourcing brings “Scarista” most of the money — 60% of the proceeds it receives from processing applications to the MFI. Another 30% earns the individual scoring models for MFI’s, 10% in consulting services. In 2014, “Skorista” handled 120 thousand applications, in 2015, 350 million (revenue amounted to 4.5 million rubles). And for one April 2016 — already 60 thousand applications. The average cost of processing an application — RUB 20 total — 1.2 million rubles. per month. In the same April, the company sold two individual scoring models; the average check — 300 thousand rbl. the Monthly revenue of the project amounted to 1.8 million rubles. Profits — about 10% of revenue, says Weichman.

Side view

INVESTORS

“Skorista” will work with the banks and outside the CIS”

“>

Pavel Nikonov, the investment Manager FRIA

The decision “Christy” unique to small microfinance organizations. The alternative is to develop their own scoring models (difficult and expensive) or buying decisions on banking scoring (expensive, ill-suited for MFIs).

For effective work in the market of microloans MFIs the necessary tools of analysis of the creditworthiness of borrowers than from a Bank. This is due to the fact that the segment of MFI-borrowers has its specificity (lack of credit history, overdue loans, etc.), does not allow to use standard banking techniques and products in assessing borrowers. Banking scoring products (for example, market leader FICO) cost a lot of money, operate on the principle of “black box” (i.e. the decision-making process occurs within a closed system) and do not solve the problems of MFIs. Thus, the market no specialized software solutions for the assessment of MFI-borrowers.

FDII believes that in the future “Skorista” will work not only with MFIs and banks, as well as outside the CIS.

“Skorista” to generate income”

“>

Igor Pesin, partner, investment Director of the Fund Life Sreda

“Skorista” virtually the only company that could immediately generate income, which is extremely rare for the segment of Big Data. In 2016 “Skorista” indicators significantly superior to the forecasts that were laid out in the business plan, both operational and financial. Small deviations from the target rate were in the middle of 2015 in connection with the General recession in the market.

The strength of the company is its leader and CEO, Maria Weichman: it is currently the best Russian expert in the field of Big Data and scoring and combines this knowledge with entrepreneurial talent and effective management; she has experienced and coordinated team of developers and mathematicians.

A weakness, in my opinion, is not sufficiently built commercial function: sales Department was reformed several times due to lack of results and only in the last six months there has been significant progress.

“Skorista” the longest works in the market of alternative scoring. This is extremely important because the market scoring is not possible to go fast — will require thousands of hours of work and a huge amount of data to build effective mathematical models. I assume this year “Skorista” will be released at least on one foreign market.

COMPETITOR

“Considered the company as one of the benchmarks

“>

Sergey Vesovic, risk Director group of companies Bystrodengi

With Skorista” we know, however, test their model was not performed, because it was originally planned to use its own design. But when studying market scoring models considered her as one of the benchmarks for assessing the quality of internal developments.

EXPERT

“Competitors “Christy” have two important elements”

“>

Yaroslav Kabakov, Director Training Corporation REU them. G. V. Plekhanov

“Skorista” integrated into all major accounting systems such as HES AFK, Terrasoft, 1C, Micfinsystems, etc. In the market a cloud-based scoring systems can distinguish competing services offered by NBCH and Equifax, but they are unlike “Christy” have expertise in more “good” borrowers (such as in MFI a few) and have two important components: online feedback on loans and extensive (over 5000) of the library of variables that help to predict the repayment of loans.

The Numbers “Christy”

€100 thousand received “Skorista” from their first investor in February 2014. In the summer of the same year, the venture Fund Life Sreda invested in the service was $400 thousand

7 thousand variables currently used in the “Scarista” to assess the borrower. Originally they were 14 times less

60 regular customers at the moment are using the service

1.8 million RUB made up the revenue service in April 2016. Profit equal to 10% of this amount

Approximately 300 thousand RUB it in “Scarista” development of individual scoring models for MFI

43% OOO Skorista” belong to the founder of the company of Mary, Weigman. Another 40% belongs to “Smartmarket”, which is owned by Vladislav Solodkiy, managing Life Sreda. 11% are owned by Alex Red, 4% — Ivan Tretyakov and 2% FRII

Source: company data, to EGRUL