How to read per visit emissions and why industry and traffic type change the picture

Per visit emissions describe the greenhouse gas associated with a single user session or single page view depending on your definition. That single number bundles work done across networks, servers and the user device. It is useful because it links technical performance to business metrics such as sessions and conversions. It can also be misleading unless you understand the main drivers and the role of traffic type and industry patterns.

What a per visit emission number usually includes

Most practical per visit measures combine a small set of components. They are network transfer energy, server infrastructure energy for hosting and CDN, and device energy consumed rendering the page. Many teams also add embodied emissions from hardware procurement or emissions from downstream services when a visit triggers heavier backend work. Which components you include should be explicit in any benchmark.

Key variables that drive per visit emissions

Three categories of variables explain most variation in per visit emissions. Understand them before comparing with any benchmark.

1. Page and session weight. Total bytes transferred per page and number of pages per session. A marketing landing page and a product search session differ by an order of magnitude.
2. Backend compute and third party work. Personalized queries, recomputed recommendations, analytics and ads cause server compute and extra network transfers. Third party scripts can add both network transfer and device work.
3. Energy intensity of infrastructure and devices. Hosting location grid carbon intensity and the efficiency of client devices matter. Traffic delivered across a grid with high carbon intensity will yield higher emissions for the same energy use.

Benchmarks by industry: a practical classification

Exact gram per visit numbers vary by measurement method and regional electricity mix. The most robust way to benchmark is to group sites by typical content and interaction patterns and use ranges that reflect relative differences. Below are four practical bands with typical industries and the main reasons for their placement.

• Low relative per visit emissionsMarketing brochure sites, documentation hubs and small blogs. These sites often serve mostly static content, have low page counts per session and can use aggressive caching. The dominant work is delivering static assets.
• Moderate relative per visit emissionsNews sites, B2B product sites and corporate portals. Typical sessions include multiple articles or pages, images of moderate size and some personalization or third party embeds.
• High relative per visit emissionsRetail and marketplace sites, multimedia publishers and SaaS apps with dashboards. Sessions tend to include image heavy pages, repeated API calls, user specific content and heavier third party stacks.
• Very high relative per visit emissionsStreaming platforms and data intensive tools that deliver long form video, complex client side rendering or large dataset downloads. Single visits can trigger long transfers and sustained device compute.

Use these bands to set target percentiles rather than a single magic number. For example aim to be in the lower quartile among peers within your industry band before promising absolute gram reductions.

Benchmarks by traffic type: why acquisition channel changes per visit impact

Traffic source influences both the type of pages users see and the length of sessions. That changes the per visit emissions profile and the carbon per conversion. Consider these typical patterns.

1. Organic searchOften lands on a specific content page. If SEO redirects traffic to targeted landing pages, per visit emissions are usually closer to single page view numbers. Organic traffic tends to yield lower per visit emissions than paid search when users bounce after finding quick answers.
2. Paid searchOften sends users to rich conversion focused landing pages with tracking and experiment scripts. Paid channels can therefore have higher per visit emissions and a higher carbon per conversion unless landing pages are optimised.
3. SocialTraffic from social platforms often includes many short sessions with media heavy previews. Social campaigns that use video or rich embeds tend to increase per visit emissions compared with text only posts.
4. EmailEmail campaigns usually land users on targeted pages with fewer third party trackers. Per visit emissions from email traffic can be moderate provided landing pages are lean and caching is effective.
5. Referral and affiliatesBehavior varies widely. Referral links from blogs or partner sites may produce long sessions if users explore multiple pages. Affiliate tracking scripts can increase device work and network transfer.

Because acquisition channels differ in conversion intent, a useful KPI is carbon per conversion or carbon per key action rather than carbon per visit alone.

How to measure per visit emissions reliably for benchmarking

Benchmarks are only valuable if measurement is consistent. The following steps produce repeatable per visit emissions that you can compare across sites and channels.

1. Decide the scopeExplicitly record whether your per visit number includes only network and device energy or whether it also includes server compute, third party cloud functions and embodied emissions. Keep the same scope when comparing benchmarks.
2. Collect page weight and session dataUse client side real user monitoring to capture bytes transferred, number of requests, pages per session and events that trigger server compute. Supplement with synthetic tests that record full page loads from major regions and devices.
3. Map energy per byte and device powerChoose an energy intensity for network transfer and an average device power profile. Document the sources and update them periodically. When possible run sensitivity tests with multiple assumptions to show range of emissions.
4. Apply regional grid intensityConvert energy to emissions using the grid carbon intensity for the hosting and major user regions. Use time of day weighting for high precision benchmarks.
5. Aggregate by traffic source and page typeGroup measured visits by acquisition channel and by primary page template. Produce median and 75th percentile values rather than single averages to limit the effect of outliers.

Example calculation using a single page visit

The following is an illustrative example to show the calculation steps. These numbers are hypothetical and only serve to explain the method.

1. Measured page transfer: 2.5 MB of data served to the device.
2. Assumed network energy intensity: 0.06 Wh per MB. Multiply 2.5 MB by 0.06 Wh per MB to get 0.15 Wh network energy.
3. Client device incremental energy to render the page: 0.4 Wh. Add to network energy to get 0.55 Wh total energy per visit.
4. Hosting and CDN energy allocated to this visit: 0.05 Wh. New total 0.6 Wh per visit.
5. Grid carbon intensity for the region: 0.4 kgCO2e per kWh equals 0.0004 kgCO2e per Wh. Multiply 0.6 Wh by 0.0004 kgCO2e per Wh to get 0.00024 kgCO2e or 0.24 gCO2e per visit.

Label this result as an illustrative metric. Replace the assumptions with your measured bytes, device energy models and regional grid intensities to produce a real benchmark.

How to build comparable benchmarks across industries and channels

To create a comparable internal benchmark set use consistent measurement scope and report these elements for each cohort.

• Cohort definitionDefine cohorts by industry vertical, by page template and by traffic source. For example B2B product detail pages reached via organic search form a cohort.
• Report distribution metricsPublish median, 25th and 75th percentile per visit emissions for each cohort and the sample size of visits measured.
• Include conversion ratiosReport carbon per conversion in addition to carbon per visit so marketing and product teams can compare channels on a business relevant basis.

Decision rules for action and prioritisation

Benchmarks are useful only if they guide action. Use these decision thresholds when choosing where to focus optimisation effort.

1. If a cohort is above the 75th percentile among peers in the same industry band and traffic source, treat it as high priority for optimisation.
2. If carbon per conversion is higher than other channels for the same campaign, invest in landing page weight reduction and tag audit work before increasing acquisition spend.
3. If a small percentage of visits account for most of the emissions, isolate the heavy templates and treat them as separate projects rather than sweeping sitewide changes.

Practical optimisation levers that move per visit emissions down

Four levers return the largest improvements for most sites. Apply them in that order to get the best cost to impact ratio.

1. Reduce bytes transferred for high traffic pagesFocus on images, fonts and large script bundles. Responsive images and proper caching usually give big wins.
2. Audit and control third party scriptsRemove or gate analytics, ads and social embeds that do not produce measurable business value per visit.
3. Shift compute from synchronous to asynchronousIf visits trigger heavy backend work, consider batch processing, on demand APIs and cache warming to reduce per visit server compute.
4. Optimise hosting and deliveryUse CDN caching, origin shielding and right sized instances. Prefer regions with lower grid carbon intensity for predictable workloads where latency constraints allow.

Reporting practices and communication to stakeholders

When you present per visit benchmarks to leadership or clients make the scope and assumptions explicit. Report both carbon per visit and carbon per conversion. Show the distribution of visits and include sensitivity bands to reflect uncertainty in device and network energy models. That transparency prevents misleading comparisons and supports credible target setting.

Benchmarks are a decision tool. Use them to prioritise technical fixes that reduce real world emissions without harming user experience or business outcomes.