Carbon Aware E Commerce: Real-Time, Region-Specific Measurement and Abatement of Web Emissions

E-commerce pages are among the heaviest on the web, with images, video, third-party scripts, fonts, and client-side JavaScript driving network transfer and computation. Prior work often reports a single "average grams CO₂e per page view," but such figures vary by methodology and have recently been revised downward (e.g., Website Carbon v4 cites ≈approximately 0.36 g CO₂e/page view globally), masking large differences by geography, device, and content mix. To provide decision-useful evidence for online retail, we introduce a carbon-aware measurement and abatement framework that combines real user monitoring (RUM) of page bytes and execution with the Green Software Foundation's Software Carbon Intensity (SCI) methodology and hourly grid intensity data. We define functional units aligned to business value in terms of CO₂e per session, per order, and per dollar of gross merchandise value, and estimate the contributions of media, scripts, CSS, fonts, and third-party integrations. We then evaluate interventions—next gen image formats, JavaScript consolidation/splitting, critical CSS loading, and font subsetting/WOFF2 alongside a carbon adaptive mode that dynamically downshifts high-impact assets when local grid carbon intensity is high. We report effects on emissions, Cumulative Layout Shift, and conversion. Our results show how e-commerce sites can reduce operational emissions while maintaining commercial performance, and we offer a reproducible protocol and open data to support adoption.