Methodology

Stackwatch builds a longitudinal dataset of what public Indian D2C storefronts actually run.

What we measure

• Platform signals (Shopify / Woo / Magento / custom) from headers and public assets
• Theme name/version when exposed (e.g. Shopify.theme)
• App/script hosts matched to a maintained signature list
• TTFB and HTML size from the homepage fetch
• Lighthouse performance on a nightly sample (~25 stores), not the full batch

What we never store

• Product copy, prices, images, reviews text, customer data
• Anything behind login or paywall

Sample sizes

Current dataset n=36. Rates, rankings, and trends publish only when n≥30. Under that floor we show raw counts or “insufficient data.”

Known limitations

• Homepage-only — apps loaded only on PDP/checkout may be missed
• Signature rot when vendors change CDNs (we alert on >40% one-night drops)
• Performance is sampled; LCP is not full-population
• No causal claims — associations only