A web page transfers data and uses servers, networks and devices. Cutting unnecessary images, scripts and requests can improve speed, user experience and estimated energy consumption all at once. To make that analysis accessible we built Green Web Audit, a free tool that brings together technical indicators normally scattered across different services.
The report does not hand out a "green" certificate and it is no substitute for a certified energy audit. It is there to show you where a page carries weight, what data is available on real-world performance, which infrastructure answers the request and which changes deserve a technical review.
What Green Web Audit analyses
Enter a URL and the tool builds an overview that can include:
- a Lighthouse test under lab conditions;
- aggregated CrUX data, where available for the URL or the domain;
- total weight and its split across images, JavaScript, CSS, fonts and HTML;
- an estimate of the CO₂ associated with transferring the page;
- information on server, IP, DNS and visible infrastructure;
- the presence of a sitemap and other public technical signals;
- priority actions to verify on the real project.
Some data may not be available. CrUX, for example, needs a sufficient volume of real experiences collected from Chrome. A missing figure does not mean the site is fast or slow: it only means that sample cannot be shown.
How to read the CO₂ estimate
A web page's CO₂ is not measured directly by the browser. It is estimated from the data transferred and from models that connect traffic, energy and emissions. The value depends on assumptions about caching, repeat visits, the network, data centres and the energy mix.
That is why the estimate is most useful as a comparative tool:
- comparing homepage, category and product pages;
- measuring before and after an optimisation;
- spotting templates that transfer far more resources than the rest;
- estimating the effect of traffic on particularly heavy pages;
- setting an internal threshold for new components and campaigns.
It should not be used as a certified figure in a sustainability report without an adequate methodology, scope and verification. In the article on CO₂, performance and eCommerce we go deeper into the relationship between page weight, traffic and sales processes.
Lighthouse and CrUX: lab and real users
Lighthouse runs a controlled test and returns repeatable metrics and diagnostics. It is useful for identifying blocking JavaScript, unsized images, main-thread work and other technical opportunities. The result can vary between runs and does not represent every user.
CrUX, by contrast, collects aggregated data from the real experiences of eligible Chrome users. Field metrics reflect the devices, networks and pages people actually use, but on their own they do not explain the cause of a problem.
Reading lab and field data together avoids two mistakes: chasing a perfect score with no real benefit, or ignoring a problem users are living with because the local test looks fast.
Resource weight tells you where to start
The report separates the main families of resources. That breakdown helps set priorities:
- Images: format, compression, responsive sizes and lazy loading.
- Video: poster images, deferred loading and the delivery platform.
- JavaScript: bundles, unused libraries, marketing tags and third-party code.
- CSS: unused styles, duplication and critical loading.
- Fonts: variants, subsets, preload and display strategy.
- HTML: repeated components, excessive markup and injected content.
Images are often a high-return area for eCommerce. The guide to WebP, AVIF and srcset shows how to cut transfer without serving files that are too small or blurry.
Hosting, DNS and responsibility
Knowing the provider, IP and DNS helps you get your bearings, but it is not enough to judge hosting quality. The same provider can offer very different configurations; a shared IP is not automatically a bad sign; server location has to be read alongside your audience, CDN, caching and legal requirements.
A full technical audit also checks response times, availability, backups, monitoring, updates, incident handling and access management. Green Web Audit shows what can be observed from the outside and helps you put sharper questions to your supplier.
Performance, SEO and conversions
A faster site does not automatically win a better position. Google uses many signals and content relevance remains central. That said, a slow page can worsen the experience, increase abandonment, make campaigns more expensive and hinder crawling on very large sites.
For an eCommerce site, performance has to be checked on the real templates: homepage, category, search, product page, basket and checkout. Optimising only the lightest page produces a pleasant report and very little operational insight.
A practical roadmap after the test
- Test several representative URLs, not just the homepage.
- Save the initial results and note the date, device and conditions.
- Rank the work by bytes, time and user impact.
- Fix the resources shared by every template first.
- Measure again after each group of changes.
- Check that analytics, consent, checkout and commercial functions still work.
- Monitor real-world data and regressions over time.
This sequence avoids decorative optimisation. A change is worth making when it cuts waste without breaking functionality, measurement or visual quality.
Frequently asked questions
Is a page's CO₂ figure an exact measurement?
No. It is an estimate based on transferred data and energy assumptions. It is useful for comparing pages and tracking improvements, not as a certified measure of corporate emissions.
Do Lighthouse and CrUX show the same data?
No. Lighthouse runs a lab test under simulated conditions; CrUX collects aggregated data from real Chrome users when the site has enough volume.
Does a faster site automatically rank better?
No. Performance improves experience and conversion and forms part of the technical signals, but it does not replace relevant content, structure, authority and the ability to satisfy the search.
What is the first thing to do to make a page lighter?
Start by measuring which resources weigh the most. Images, video, fonts and third-party scripts often offer the biggest savings, but the priority depends on the individual page.