TechPrufer
Get started
Shopify
Shopify logoShopify · Feature trail

Quantile Regression for Merchant Funding

2 posts · Jul 2017 – Nov 2021

What this trail covers

Shopify has evolved its capabilities for assessing merchant eligibility for cash advances by applying machine learning techniques, specifically quantile regression. This allows for more nuanced predictions of future sales and risk assessment, ensuring that advances are structured for a high probability of repayment. The system aims to provide funding that balances risk for both merchants and Shopify.

All Shopify trails
  1. Using Propensity Score Matching to Uncover Shopify Capital’s Effect on Business Growth - ShopifyThis post details the application of Propensity Score Matching (PSM) to evaluate the impact of Shopify Capital on merchant business growth. It explains the research question, the challenges of conducting a true A/B test, and the methodology used to create comparable treatment (US first-time Capital adopters) and control (Canadian shops not yet eligible for Capital) groups. The post outlines the selection of covariates, the use of caliper matching, and the techniques for assessing matching quality (standardized mean differences, visual diagnostics, and variance ratios) to validate the causal inference.
  2. How Shopify Capital Uses Quantile Regression To Help Merchants Succeed - ShopifyThis post details the application of quantile regression for predicting merchant sales and assessing cash advance eligibility. It explains how quantile regression differs from standard regression by fitting conditional quantiles rather than just the mean, providing a measure of confidence in predictions. The post uses examples of linear regression and wine quality prediction to illustrate the concept and its practical application in managing risk for both merchants and Shopify.