Website speed
How to test your website speed without getting lost in the numbers
Key takeaway
Use Google PageSpeed Insights to test a public page on mobile and desktop, then try the real customer journey on a phone. A simulated score and actual visitor measurements answer different questions. Record the page, device and date, and compare the same conditions after making an improvement.
In this guide
Start with one important public page
Open Google PageSpeed Insights, paste the full page address and run the analysis. Look at both mobile and desktop. Start with a page that brings enquiries, then repeat the exercise for a typical service page and the contact or booking page.
Save the report address or a screenshot with the date. Include the exact page you tested. A report for the homepage cannot tell you everything about a different page with a gallery, map or booking tool.
Find out where a slow page is waiting
Sources: Google PageSpeed Insights testing tool (opens in a new tab)
Separate a test from actual visitors
The Lighthouse part of PageSpeed Insights runs a controlled test. The real-user section, when available, draws on Chrome visitor experience over the previous 28 days. A small or new website may not have enough eligible data for that section.
A missing real-user result is not a failed speed test. Use the available diagnostics and real checks while data develops. Also check whether the report describes that exact page or the wider website, because those are different views.
Sources: Google: what PageSpeed Insights measures (opens in a new tab)
Read the useful finding before the score
Look for what kept the main content waiting, whether a tap received a prompt response and whether the page moved unexpectedly. The report’s suggested improvements give a developer a starting point. You do not need to understand every abbreviation before asking for help.
A simulated score can change with test conditions. Compare like with like and use more than one run when deciding whether a change helped. The separate SEO audit score checks selected technical items; it is not a forecast of your Google position.
Turn the findings into speed improvements
Sources: Google: why test and real-visitor results can differ (opens in a new tab)
Try what a customer actually does
On a phone, open the page over mobile data. Read the service information, use the navigation and try the enquiry or booking path. If you need to send a test enquiry, agree a clearly labelled test with the business first.
A report cannot decide whether the right phone number is visible or the wording explains your service. Record those problems alongside speed findings. Improving the customer experience is more useful than collecting a better score in isolation.
Bring the result into a hosting discussion
Send us the page address and tell us where the wait appears. We assess the website before confirming whether moving hosting, preparing smaller images or modernising the website is the suitable route. Any separate work is agreed first.
Both Lightning Web Hosting plans include AWS hosting and CloudFront CDN. Thunder is NZ$55 + GST per website a month; Lightning is NZ$99 + GST and adds everyday Arc updates. Hosting starts billing when your website is live with us.