What is an alternative credit score and how does it work?
Insights by Finvero
For decades, the decision to lend money relied almost entirely on the history a person or business has with Credit Information Companies (SIC, for its Spanish acronym), better known as the credit bureau. The problem is that millions of people in Mexico and Latin America have little or no credit history: young people, freelancers, entrepreneurs, or anyone who has never had a credit card. To the traditional model, they are invisible.
The alternative credit score was created to close that blind spot: a risk rating that, in addition to traditional history, analyzes other sources of information to accurately estimate the probability that someone will repay their loan.
What is an alternative credit score?
It is a scoring model that combines traditional data with alternative data and machine learning techniques to predict repayment behavior. The result is a score the lender uses to decide whether to approve, decline, or adjust the offer: amount, term, or rate.
Unlike a traditional score, it does not depend on the person having had credit before. That is why it is especially useful for the no-hit segment, people who do not appear in the credit bureau, and the thin-file segment, people with very little information.
What data does an alternative score use?
Sources vary by country and regulation, but they typically include:
Digital behavior: how the person interacts with the application, their device, and their digital channels.
Transactional and income data: cash flows, recurring payments, and income stability.
Employment and social security history: continuity and formality of employment.
Utility and phone bill payments: signals of everyday financial responsibility.
Identity validation against official sources: in Mexico, with RENAPO and the INE voter registry.
At Finvero, the model for the no-history segment analyzes more than 700 alternative variables, always with the user's consent and in compliance with Mexico's Federal Law on the Protection of Personal Data Held by Private Parties (LFPDPPP).
How is it different from the credit bureau?
Coverage: the credit bureau only scores people who already have a history; the alternative score also evaluates those who don't.
Timeliness: credit history looks backward; alternative data reflects the applicant's current situation.
A complement, not a replacement: when the applicant does have a history, the alternative score enriches it with more context for better decisions.
How do you measure whether an alternative score is good?
Risk teams use predictive power metrics such as AUC, which measures how well the model ranks good and bad payers, and KS, which measures the separation between the two groups. For reference, Finvero's Alternative Risk Score reports an AUC of 87.7% and a KS of 65.5%.
It also matters that the model is explainable and auditable: every decision should retain the variables behind it, which is key for regulated institutions.
Benefits for lenders, banks, and fintechs
More approvals with the same risk appetite: you find good payers the traditional model used to decline.
Lower delinquency: you detect risk that credit history doesn't show.
Decisions in seconds: the score is queried via API within the origination flow.
Profitable financial inclusion: you serve underserved segments without compromising your portfolio.
Conclusion
An alternative credit score lets you see the people and businesses the credit bureau can't, and make better-informed decisions about those it does cover. For any institution that wants to grow its portfolio without growing delinquency, it is one of the most impactful tools available. See how it works in the Alternative Risk Score and in our solution for the NoHit segment.
Frequently asked questions
Does an alternative score replace the credit bureau?
No. It complements it: it evaluates people with no history and, when a history exists, enriches the decision with more context.
Is it legal to use alternative data to assess credit in Mexico?
Yes, as long as you have the user's consent and comply with the Federal Law on the Protection of Personal Data Held by Private Parties (LFPDPPP).
How accurate is an alternative score?
It depends on the model. Accuracy is measured with indicators such as AUC and KS; Finvero's Alternative Risk Score reports an AUC of 87.7% and a KS of 65.5%.