Our method
How we measure New Zealand’s websites, step by step. Every month our bot opens each website’s pages on a phone and a laptop and counts everything they download.
Same pages for all.
Each website is measured on its homepage and the same four kinds of page as the rest of its industry. For a bank: mortgages, credit cards, everyday accounts and contact.
A first-time visitor.
We open each page in Chrome, as a phone and as a laptop, with nothing saved, from New Zealand. Every file it downloads counts, whoever it belongs to.
Two ways in.
First visit: each page opened with nothing saved, as if from a search. Returning: each page reached by a click from the page before, with the website’s files already saved.
When we stop.
When the page has loaded and nothing more has downloaded for three seconds, we stop counting. We never wait longer than 30 seconds, however busy the page is.
See the pages for every industry
The figures
- Every grade comes from one number: a typical page view. For each page we measure, we take three parts first visit and one part returning, because most people arrive fresh and some come back. We take the middle page on a phone and on a laptop, and average the two.
- First visit: a page opened with nothing saved, as if you had found it on Google.
- Returning: a page reached by clicking from another page on the same website, with some of its files already saved.
- Phone and laptop count equally. All weights are in megabytes (MB).
Savings
We find each website’s biggest savings by making the change to a copy of its files: smaller images, compressed text, trimmed fonts, a still image in place of video, code that never ran, and files the next pages download again. We show the five biggest, worked out on a phone and a laptop. Other companies’ files are named, never changed.
“A smaller website does not switch off a data centre. It does mean less work for each device that opens it.”
- Every file a page downloads on a first visit.
- What is new on a later page.
- Files from other companies, named where we can.
- Your delivery network’s files.
- A share of the energy used by data centres, networks and visitors’ devices, and of making them.
- Anything that loads only after someone clicks or accepts cookies.
- What your own servers actually do, which we cannot see from outside.
- Video, which we weigh and show on its own.
- Security checks that stop a visit, which we never measure.
Where we could be wrong
- We make up to three visits on each device, from one place, at one connection speed. Visitors on slower connections or other devices may see more or less.
- Some websites load different things on different visits. We use the middle of our visits and show the range.
- Some websites keep downloading for as long as a page is open. We stop at the point described above.
- Savings are estimates, worked out on copies of a website’s files. The real change may save more or less.
- A few websites are measured manually by one of our team, by the same method, and their profiles say so.
| What | Source | Updated |
|---|---|---|
| New Zealand's grid emissions | Ministry for the Environment, Measuring Emissions Catalogue 2026 v2: purchased electricity, 2025 annual average, 0.0786625 kg CO2e per kWh, for visitors' devices and networks; the world average, 494 g per kWh, for data centres | next review May 2027 |
| From data to carbon | Sustainable Web Design model, version 4, through CO2.js 0.19.0: 90.5 g CO2e per gigabyte | next review May 2027 |
| Grade lines | Sustainable Web Design digital carbon rating, version 4 (A+ to F); G and H are ours | fixed |
| Real-visitor figures | Chrome UX Report, Google, CC BY 4.0 | monthly |
| Who other companies' files belong to | the open third-party-web dataset | as published |