Website performance and Core Web Vitals metrics explained

Why website performance matters to the person paying, not just the person building

Speed is the attribute this site argues hardest on in its own copy, which is a reason to be more careful here rather than less. These pages explain the performance metrics and their limits, and they do not quote figures for what a speed improvement is worth in enquiries. Those figures circulate widely, they come from large sites in other markets, and transplanting them onto a small Australian business would be exactly the unsourced-statistic behaviour criticised elsewhere on this site.

What can be said accurately is what each measurement measures, what moves it, and how to find out where your own site stands.

What the performance pages cover: Core Web Vitals, load speed and page weight

Start with the measurements themselves — Core Web Vitals explained — then each in turn: Largest Contentful Paint (LCP, the vitals metric for how many seconds the main content takes to appear), Interaction to Next Paint and Cumulative Layout Shift.

Then the causes, ranked by how often each turns out to be the one that matters: why websites are slow, and the single largest contributor on most small-business sites, images, formats and why they dominate page weight.

Then the infrastructure half, which is smaller than it is usually claimed to be: what hosting does and does not do for speed, Australian hosting and latency, and caching and CDNs.

Finally, how to test your own site with a free tool such as GTmetrix or PageSpeed Insights, which takes about twenty minutes with no technical background, and which is worth doing before any supplier quotes on fixing it.

The distinction that matters most: server response time vs page load

Two different problems get called “slow”. One is the server taking a long time to answer at all, which hosting controls completely — this shows up as a delay before anything loads at all, measured in seconds. The other is the page being heavy once it arrives, which hosting cannot touch. They have different symptoms, different fixes and very different costs, and quotes routinely address one while the site suffers from the other. Telling them apart is the single most useful thing in this group.

What performance figures this site will not quote, and the honest link to conversion

It will not tell you what a one-second improvement is worth in enquiries, or attach a specific conversion rate lift to a speed fix.

Those figures circulate widely and they come from very large sites in other markets, measured under conditions that are not stated alongside the number. Transplanting them onto an Australian small business would be exactly the unsourced-statistic behaviour criticised elsewhere on this site. What can be said accurately is the direction, not the size: a page that loads slowly loses visitors before they see anything, which shows up as a higher bounce rate, and a visitor who bounces cannot convert. That is a real and well-established link between speed and conversion rates — it is the specific multiplier that gets invented, not the relationship itself. Knowing what moves each measurement, and how to find out where your own site stands, is more useful anyway, because it is actionable.

GTmetrix, PageSpeed and other page load speed test tools

GTmetrix and PageSpeed Insights are the two free page load speed tools most owners reach for first, and both report the same Core Web Vitals metrics — Largest Contentful Paint, Cumulative Layout Shift and Interaction to Next Paint — from a real page load rather than a lab estimate. Running one before a supplier quotes on a speed fix turns a vague complaint about load time into a specific number a fix can be measured against.

Where the effort actually pays

Performance work has a strongly uneven return, and knowing the order saves money.

1. Images first, almost always

On a typical small-business site, images outweigh every other asset combined, and the fix is largely mechanical — correct formats, correct dimensions, lazy loading below the fold. It is the cheapest large improvement available and it rarely requires touching the design.

2. Then whatever is blocking the first render

Fonts that load before text can appear, third-party scripts that sit in the critical path, a carousel nobody asked for. These are decisions rather than defects, and reversing them is usually a conversation about whether the element earns its cost.

3. Then the server

This is where most people start, and where the least is usually available. Hosting controls how long the machine takes to answer. It cannot make a heavy page light. Moving host to fix a page-weight problem is the most common wasted spend in this area.

4. Last, and only once the first three are done, the deeper build questions

How the site is rendered, what runs on the client, whether a framework is earning its size — these are worth investigating only after images, render-blocking assets and hosting have already been ruled out.

Where performance sits inside web design as a discipline

For where performance sits inside web design as a discipline, start with web design in Australia. A site’s speed is one line item among the many parts a build has to get right, and it is rarely worth chasing in isolation from the rest of the project.