How to calculate emissions per visit

Emissions per visit are the product of energy used to deliver and render a visit multiplied by the carbon intensity of the electricity that powered that energy. Breaking the problem into components makes estimates reproducible and comparable across sites and traffic types.

Core components

Estimate three energy components for a typical session or pageview and sum them.

• Network energy required to transfer bytes from origin to user via CDN and ISP infrastructure.
• Server and CDN energy consumed at the origin and edge to serve the request and any on‑the‑fly work.
• Device energy used by the user device to decode, render and display the page and to run any client side scripts.

After you have a total energy per visit in kilowatt hours you convert to emissions by multiplying by a grid carbon intensity in grams of CO2 per kilowatt hour. Use a country or region specific number where possible. If you report a single number for an international audience, show sensitivity to at least two or three grid intensities so readers understand the range.

Reusable formula (single visit)

Emissions (gCO2) = (E_network + E_server + E_device) in kWh × grid intensity (gCO2/kWh)

Practical measurement inputs and example assumptions

Most teams can measure page weight and session structure from analytics and lab tools. The only parts that typically need external reference values are energy intensity per gigabyte for network and per‑request server energy. Below are conservative example assumptions that make calculations reproducible. Treat these as an explicit example scenario rather than as universal constants. Replace any value with measurements or vendor data where available.

• Measured inputs you should gather: median page size in megabytes, median pages per session, percent of sessions on mobile versus desktop, average video minutes per session, and average API calls per session.
• Example energy intensities (illustrative): network energy intensity 0.06 kWh per GB, server and CDN combined 0.01 kWh per GB served, smartphone render energy 0.2 Wh per pageview, desktop render energy 1.5 Wh per pageview. Document any substitution or measurement you use.
• Grid intensity examples: low 50 gCO2/kWh, medium 250 gCO2/kWh, high 500 gCO2/kWh. Use a local figure from an official source for final reporting.

These example intensity values are drawn from public literature and technical reports and are intended to illustrate the calculation pattern. Replace them with better local or measured numbers when possible. See the sources section for references to further reading on typical energy intensities and grid data.

How traffic type changes per‑visit emissions

Traffic type alters emissions per visit primarily by changing three variables: bytes transferred per session, client processing work, and the likelihood of video or long polling. The same pageweight on mobile and desktop may generate different emissions because device energy for rendering and display differs by device class and screen brightness.

Single page marketing visits

Landing pages are usually low in page count per session, but can be heavy with images, fonts and tracking scripts. Because sessions are short, network bytes per session tend to dominate. Optimizing image formats and limiting third party scripts reduces per‑visit emissions quickly.

Multi page content sites

News and editorial sites combine higher pages per session with often moderate page sizes. Repeated navigation magnifies server and network reuse benefits from caching. Reducing repeated payloads and enabling shared caches lowers per‑visit emissions significantly for returning readers.

Ecommerce sites

Ecommerce sessions often include many pageviews, product image loads, and dynamic requests during browsing and checkout. Personalization and client side rendering can raise device energy. Emissions per completed purchase are often a more useful business metric than emissions per visit because conversion funnels differ across merchants.

Video and streaming

Sites where video is central shift the dominant term to duration‑based transfer. Each minute of video drives large additional network energy compared with static pages. Use minutes streamed per session and the video bitrate to calculate extra GB and thus the incremental energy and CO2.

Example benchmark calculations by industry and traffic mix

The figures below are example estimates produced with the reproducible assumptions listed earlier. They are not authoritative universal benchmarks. Use them to understand relative differences and to test your own numbers by replacing the assumptions with measured inputs.

Assumptions used for all examples

• Network energy intensity 0.06 kWh per GB.
• Server and CDN energy 0.01 kWh per GB.
• Smartphone render energy 0.2 Wh per pageview. Desktop render energy 1.5 Wh per pageview.
• Example grid intensities shown as low 50 gCO2/kWh, medium 250 gCO2/kWh, high 500 gCO2/kWh.

1. Marketing landing page (single page session)

Typical measured inputs: page weight 1.5 MB, pages per session 1, 80 percent mobile sessions.

Network plus server energy = (1.5 MB / 1024) × (0.06 + 0.01) kWh per GB = approximately 0.0001 kWh per visit.

Device energy = 0.8 × 0.2 Wh + 0.2 × 1.5 Wh = 0.16 Wh + 0.3 Wh = 0.46 Wh = 0.00046 kWh.

Total energy ≈ 0.00056 kWh. Emissions at medium grid intensity 250 gCO2/kWh ≈ 0.14 gCO2 per visit. At low intensity 50 gCO2/kWh ≈ 0.028 gCO2 per visit. At high intensity 500 gCO2/kWh ≈ 0.28 gCO2 per visit.

2. News or content site (multi page session)

Typical measured inputs: page weight 2.0 MB, pages per session 4, 70 percent mobile sessions.

Network plus server energy = (2.0 × 4 MB / 1024) × 0.07 kWh per GB ≈ 0.00055 kWh.

Device energy = 0.7 × (4 × 0.2 Wh) + 0.3 × (4 × 1.5 Wh) = 0.56 Wh + 1.8 Wh = 2.36 Wh = 0.00236 kWh.

Total energy ≈ 0.00291 kWh. Emissions at medium grid intensity 250 gCO2/kWh ≈ 0.73 gCO2 per visit. Low intensity 50 gCO2/kWh ≈ 0.15 gCO2. High intensity 500 gCO2/kWh ≈ 1.46 gCO2.

3. Ecommerce browsing session

Typical measured inputs: page weight average 3 MB, pages per session 6, 60 percent mobile sessions, plus three product image high resolution loads totalling 4 MB during the session.

Network plus server energy = ((3 × 6 + 4) MB / 1024) × 0.07 kWh per GB ≈ 0.0017 kWh.

Device energy = 0.6 × (6 × 0.2 Wh + image decoding 0.5 Wh) + 0.4 × (6 × 1.5 Wh + image decoding 1.5 Wh) ≈ 0.6 × (1.2 + 0.5) Wh + 0.4 × (9 + 1.5) Wh = 1.02 Wh + 4.2 Wh = 5.22 Wh = 0.00522 kWh.

Total energy ≈ 0.00692 kWh. Emissions at medium intensity 250 gCO2/kWh ≈ 1.73 gCO2 per visit. Low intensity 50 gCO2/kWh ≈ 0.35 gCO2. High intensity 500 gCO2/kWh ≈ 3.46 gCO2.

4. Video heavy session (streaming)

Typical measured inputs: initial page 1.5 MB plus 10 minutes of video at 2 Mbps. 50 percent mobile sessions.

Video GB = 2 Mbps × 600 seconds = 1.2 Gbit = 0.15 GB. Total transfer ≈ 0.15 GB + 0.0015 GB for the page. Network plus server energy ≈ 0.1515 GB × 0.07 kWh/GB ≈ 0.0106 kWh.

Device energy: smartphone display and decoding for 10 minutes might be approximately 6 Wh for mobile and 18 Wh for desktop during playback. Weighted device energy = 0.5 × 6 Wh + 0.5 × 18 Wh = 12 Wh = 0.012 kWh.

Total energy ≈ 0.0226 kWh. Emissions at medium intensity 250 gCO2/kWh ≈ 5.65 gCO2 per visit. Low intensity 50 gCO2/kWh ≈ 1.13 gCO2. High intensity 500 gCO2/kWh ≈ 11.3 gCO2.

Interpreting these example benchmarks

Three lessons are clear from the worked examples. First, per‑visit emissions depend more on session structure and device energy than on the raw page weight alone. Second, video multiplies per‑visit emissions because it adds sustained device power and large transfers. Third, local grid intensity is often the largest multiplier in converting energy to CO2, so the same site will have very different emissions per visit across countries.

How to measure per‑visit emissions on your site

1. Gather measured inputs: median page weight, pages per session, percent mobile, average video minutes per session, API call counts. Use real user monitoring and lab tools for accurate page weight split by asset type.

2. Replace the example intensities with measured values where available: CDN reports for edge energy per GB, datacenter PUE and server power for server CPU costs, and device profiling for heavy client components of your pages.

3. Use regional grid intensity for the geographic mix of your traffic. Our World in Data and national statistics provide per‑country CO2 per kWh numbers you can apply.

4. Report ranges and uncertainty. Show at least three scenarios using low, medium and high grid intensities and note variability in device mix and session structure. Avoid a single deterministic number without uncertainty bounds.

How teams should use benchmarks

Benchmarks are most useful when they inform choices that change bytes, server work or device load. Use per‑visit estimates to prioritize work that reduces repeated transfer, shrinks high frequency assets, or avoids unnecessary client computation for the most common session types. When reporting to stakeholders combine per‑visit emissions with monthly active users or conversions to produce business‑level metrics, and always document assumptions.

Next steps and measurement checklist

• Compute a baseline using your analytics and the reproducible formula above and record all assumptions.
• Run sensitivity analyses for grid intensity and device mix to produce a range rather than a single number.
• Prioritize interventions that reduce repeated bytes and heavy client work for your most common sessions. Recompute after each change to show impact.
• Publish methodology alongside any public claim so readers can reproduce and interrogate the result.

Benchmarks should guide decisions, not replace measurement. The method above scales from a single pageview to business level reporting when you multiply per‑visit estimates by measured traffic and conversion counts.