"In Argentina, the penetration of credit cards is 32%. However, only 72% of buyers use it when paying at Mercado Libre. This shows that there are a lot of people who are not buying for not have access to credit through plastic, "says Martín De los Santos, Senior VP of Mercado Crédito, the credit division of the ecommerce giant.
In this framework and taking into account that credit to the private sector in Argentina only represents 16% of GDP, Mercado Crédito set out to reach out to those excluded from the financial system with lines of financing. To do this, he developed two tools:
- On the one hand, working capital loans for SMEs that sell through Mercado Libre or use Mercado Pago to make their sales
- On the other, consumer loans of up to $ 50,000 for those who make purchases in the marketplace or pay with Mercado Pago. Thus, they access up to 12 installments even without a card
How to qualify to get it?
Unlike banks, which mainly use information available from credit bureaus (such as Veraz or Nosis) to determine who to lend to, Mercado Crédito developed its own scoring tool, with a focus on financial inclusion.
"Suppose Mario does not have access to traditional credit, like most people that banks do not give him. So, he must pay in cash or resort to informal financing, which is typically not very transparent or abusive. This is the point that we want to solve ", exemplifies iProUP De los Santos.
The Mercado Credito rating system uses some 2,400 behavioral variables to determine the credit supply and rate for each person.
:quality(70)/https://assets.iproup.com/assets/jpg/2020/06/10498.jpg?5.6.1.1)
Martín de los Santos, Senior VP of Credit Market
"We perform a segmentation both by risk profile and by the probability of adoption of the line. We know that there are people who are included in the financial system and who will not be interested in this product, for example. Regarding the risk profile , the rate goes up as the client is more risky, "he says.
There, a fundamental fact is whether the candidate uses the Mercado Libre ecosystem. If so, they study your purchases, how you navigate the site, how you interacted and paid, among other points.
De los Santos elaborates on this point. "The machine learning models we use offer the possibility of analyzing multiple variables. Artificial intelligence allows us to process large amounts of data and make better segmentations. Even much of what we analyze is combinatorial variables. For example, what did you buy and how do you buy it? financed, "he says.
The variables that Mercado Crédito analyzes are divided into six groups:
1. - Navigation: "There are some 250 internal variables that have a weight of 12% in the decision. Among them, what products do you buy, at what time, from which categories, with which computer, what is the connection speed, the questions he has done and the interaction. Although they serve, they are not as predictive, "says De los Santos
2. - Transactional: among them are the average ticket, the purchase volume, the form of payment and financing. "In that group there are 900 indicators, which have a weight of 24%," he explains.
3. - Previous loans: it is what has the greatest influence, 34%, when determining the type of loan is the behavior in previous loans, if you have already taken them. "What did he use it for, how was it financed, what percentage of the line did he use, if he fell behind on the payment," says De los Santos in relation to the 550 variables that make up that group.
4. - Collections: there are about 150 variables that have to do with collection management on previous loans. They have an 18% representation in the decision
5. - Sales history: If the credit candidates made sales in Mercado Libre, the products they commercialized and the feedback they had from customers, among other points, are also analyzed. Those 250 variables have a weight of 6% in the decision of the loan offer
6. - Credit Bureau: they are the most traditional source, which is also used by banks. They contribute 280 variables and 6% to the decision
"By having the possibility to see each person from a different angle, via technology and data, we can make credit decisions and give them a first offer of credit," says the executive.
Open sea strategy
Beyond the clients who actively participate in the unicorn's ecosystem, the manager acknowledges that there are many people who have never used the platform. "How can we include them? There we move on to the open sea strategy," synthesizes De los Santos.
"If Mario had never made a transaction with Mercado Libre and when we go to the credit bureau he has no information about him or, worse still, he has negative information, instead of excluding him, we give him a very small initial line, between $ 1,500 and $ 5,000 ", bill.
With this line, the client can make his first purchase in installments. The rate, in this case, will be closer to the maximum of 140%, in any case, De los Santos clarifies that they are not usurers, despite accusations from the Central Bank. "Obviously this is a more risky segment. Precisely, because it is negativized and does not know it, it has greater bad debt," he justifies.
However, it clarifies that it is a higher rate for a small amount and that "as you start to pay, the rate is going down and the amount is going up."
When paying with QR or buying in Mercado Libre, the user can choose to "pay with Mercado Crédito" and finance the operation on the spot in up to 12 installments
And he adds: "After 30 days, when he makes his first payment, we already began to have information regarding that group of clients. The model is already beginning to distinguish those who pay well and can be given a little more line or lower the rate. "
Regarding costs, De los Santos highlights that Mercado Crédito does not charge insurance or hidden costs. In addition, it points out that the average rate ranges between 100% and 110%, in line with the consumer credits offered by the large retail chains.
"The vast majority of customers are paying between 80% and 120%. That may be a lot for a person who is banked, but it is not for someone who never had a loan because banks do not offer it," he said.
Some results
The Mercado Crédito executive is proud of the company's financial inclusion task, since the flexibility of its own scoring system allows Mercado Crédito to approve 60% of loan applications.
"It is much more than what traditional banks approve. By giving a small first line, we can afford to approve much more. In any case, we know that there is more risk, but the customer who does not pay has a lower cost and the one who pays , we accompany you to form your credit history, "he explains.
"85% of clients are outside the metropolitan area and 80% have a score below 400 in Nosis. Banks could not lend to them by policy," he says, referring to harsh regulations.
At the same time, he points out: "55% of the buyers who used our credit product would not have been able to make the purchase if they did not have it. We are promoting and helping the consumption of durable goods by customers who did not have access to this type Shopping".
Finally, the executive enhances the loyalty that Mercado Crédito generates in customers, since 80% of the portfolio has already taken more than one loan. At the same time, he assured that the default is under control, although it increased as a consequence of the COVID-19 crisis, despite the postponement of maturities.
"We have granted more than $ 11,000 million to SMEs and consumers through 3 million loans destined to almost half a million clients. This generated a contribution of $ 1,200 million to the Treasury, due to the taxes we withhold," said De los Santos.
:quality(85)/https://assets.iproup.com/assets/jpg/2022/07/29487_landscape.jpg)
:quality(85)/https://assets.iproup.com/assets/jpg/2025/05/42802_landscape.jpg)
:quality(85)/https://assets.iproup.com/assets/jpg/2019/05/4221_landscape.jpg)
:quality(85)/https://assets.iproup.com/assets/jpg/2024/02/37904_landscape.jpg)
:quality(85)/https://assets.iproup.com/assets/jpg/2022/08/30206_landscape.jpg)
:quality(85)/https://assets.iproup.com/assets/jpg/2024/04/38334_landscape.jpg)
:quality(85)/https://assets.iproup.com/assets/jpg/2024/12/41291_landscape.jpg)
:quality(85)/https://assets.iproup.com/assets/jpg/2024/11/40935_landscape.jpg)
:quality(85)/https://assets.iproup.com/assets/jpg/2021/07/20423_landscape.jpg)
:quality(85)/https://assets.iproup.com/assets/jpg/2024/11/41064_landscape.jpg)
:quality(85)/https://assets.iproup.com/assets/jpg/2024/10/40597_landscape.jpg)
:quality(85)/https://assets.iproup.com/assets/jpg/2024/04/38332_landscape.jpg)
:quality(85)/https://assets.iproup.com/assets/jpg/2025/01/41615_landscape.jpg)
:quality(85)/https://assets.iproup.com/assets/jpg/2025/05/42795_landscape.jpg)
:quality(85)/https://assets.iproup.com/assets/jpg/2024/05/38673_landscape.jpg)