We believe in trusting relationships with our stakeholders. That’s why we’re sharing our recipe for success (without giving away the secret sauce).

What we predict

For every borrower in a partner NBFC’s book, our model estimates the probability that they fall 30, 60 or 90+ days behind on repayment. Our model is capable of predicting at multiple time horizons, so we can price an entire loan pool at once.

Our Model is our Moat

A standard credit model learns from finished stories: loans that either repaid or defaulted. But at any moment, a large share of a lender’s portfolio is still outstanding. A two-year loan disbursed eight months ago hasn’t succeeded or failed yet.

The simple fix would be to train only on loans which are old enough to have a final outcome. But that discards a large share of the most recent data. We found that this is the data that matters most. Repayment capacity moves with current economic conditions: crop prices, climate, local labour demand and funding conditions across the sector. A model that doesn’t learn from recent data can’t successfully predict tomorrow’s repayments.

Timeline of loans: older loans are fully observed; the newest loans run past today and their outcomes are not yet known. 2019 2021 2023 2025 today
Each bar is a loan, from disbursement to scheduled maturity. The dotted segments haven’t happened yet. A standard model drops these loans entirely. Ours learns from every loan, new or old.

How the model works

Picture checking in on every loan once a month and asking: did this borrower fall behind this month, or are they still on track? Our model learns from all these monthly check-ins.

Take a loan that has been paid on time for eight months and isn’t finished yet. Other models would ignore it until it ends. Ours learns from these eight good months. Nothing gets thrown away.

What goes in

Every prediction draws on layers of information, all measured at the point in time the prediction is made:

  • Repayment history

    How the borrower has paid so far: timing, gaps, partial payments, previous loans.

  • Credit bureau records

    Score, number of active loans and delinquencies with other lenders. Even whether a borrower has a bureau record at all is informative, and the model treats it that way.

  • Borrower and loan details

    Demographics, loan size, loan product and loan terms.

  • Geography and season

    Dynamic environmental data: where the borrower lives and what is happening there, month by month, from agricultural seasons to regional risk.

  • Social environment

    Microfinance runs on relationships: loan officers financially advise borrowers, and borrowers often organize in groups. Our model also tracks the overall performance of the officer and group with which each borrower is associated.

How we check ourselves

We validated our model on nearly half a million historical microloans from established Indian microfinance lenders. The model’s performance is measured by AUC: the probability that, given a borrower who is delinquent and one who isn’t, the model successfully ranks the delinquent borrower as riskier. 0.5 is a coin flip; 1.0 is perfect.

0.88AUC on held-out loans the model never saw in training
0.83AUC on loans newer than anything in training. This is the figure we plan around.

High scores in credit modelling are easy to fake by accident. Information about the outcome can slip into the inputs, or the test set can quietly resemble the training set. We run three checks on every version of the model:

Borrower-level splits. Many borrowers take several loans. We test with all of a borrower’s loans held out together, so the model can’t recognise a borrower it has already seen. The score drops by just 0.015.

Time-based splits. We train on older loans and test only on newer ones, the way the model is actually used. This is where the 0.83 comes from. The drop is concentrated in the shortest horizons, where new loans simply haven’t had time to show any history. That is the pattern of immature data, not of a model that has been cheating.

Label shuffling. We randomly reassign every loan’s outcome to a different loan and retrain. If the inputs secretly encoded the answer, the model would still score well. It doesn’t: what little it retains is fully explained by the age of the loan, a known and intended input.

What it means for our partners

When an NBFC partners with us, we score their loan books, account by account and month by month. We identify loans we are most confident in, build them into a securitised pool, and price that pool on our own expected losses. Partner NBFCs receive instant payments for loans they’ve already given out.