Finance ops · Production

Shopify Capital uses quantile regression to predict merchant sales and determine cash advance eligibility

The problem

Shopify Capital needed to accurately predict future merchant sales to offer cash advances responsibly, but standard regression only fits the average of a distribution and fails to capture prediction uncertainty — a critical gap for risk-appropriate lending decisions.

Workflow diagram · grounded in source
1
Merchant capital request
trigger
“Shopify Capital provides funding to help merchants on Shopify grow their businesses. But how does Shopify Capital award these merchant cash advances?”
2
Quantile regression sales forecast
ai_action
“For merchants who are well established and have a proven track record of making sales, our model makes predictions for their future sales that have smaller error bands. For younger merchants that are just starting out, our model makes pr…”
3
Repayment probability validation
validation
“ensuring that each advance offered has a high probability of being paid back in reasonable time”
4
Advance offer issued
output
“using the same quantile regression techniques we are able to offer merchant cash advances to Shopify merchants that make sense for their business”
5
Growth feedback into model
feedback_loop
“Those who start growing as a result of an early advance are then cycled back into the model triggering an update and allowing the model to offer them more next time”
Reported outcome

Using quantile regression, Shopify Capital can offer advances to both new and established merchants while ensuring each advance has a high probability of repayment, and merchants who grow cycle back into the model for larger future offers.

Reported metrics
Advance repayment confidencehigh probability of being paid back in reasonable time
risk to merchant and Shopifymaking sure that neither the merchant nor Shopify takes on too much risk
Reported stack
quantile regression
Source
https://shopify.engineering/how-shopify-uses-machine-learning-to-help-our-merchants-grow-their-business
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Using quantile regression, Shopify Capital can offer advances to both new and established merchants while ensuring each advance has a high probability of repayment, and merchants who grow cycle back into the model for…

What tools did this team use?

quantile regression.

What results were reported?

Advance repayment confidence: high probability of being paid back in reasonable time; risk to merchant and Shopify: making sure that neither the merchant nor Shopify takes on too much risk (source-reported, not independently verified).

How is this finance ops AI workflow structured?

Merchant capital request → Quantile regression sales forecast → Repayment probability validation → Advance offer issued → Growth feedback into model.

Related finance ops cases