Website Speed & Core Web Vitals
Fixing a slow site you already have — LCP, INP and CLS measured on real mobile connections, not a desktop lab score. Fixed price, report before any work.
Fixing a slow site you already have — LCP, INP and CLS measured on a real mobile connection rather than a desktop lab score.
Your visitors are on phones, on mobile data, often outside a city. That number is usually two to four times worse than the score you get testing on your laptop, and it's the one that decides whether they stay.
The honest bit about rankings
Most agency pages sell speed as an SEO silver bullet. It isn't, and it's worth saying so before you spend money.
Google does use Core Web Vitals as a ranking input, and it's part of the page experience signal. But the direct ranking effect isn't large, and plenty of SEOs report seeing little movement from Core Web Vitals improvements alone. Content and competition matter more.
What speed reliably changes:
Whether people stay and buy. Google's own research found the chance of a bounce rises sharply as load time goes from one second to three. Sites loading in one second convert around three times better than sites loading in five.
What this is for
- A site that 'feels slow' and you don't know why
- Failing Core Web Vitals in Search Console
- Mobile visitors leaving before the page finishes
- A site that got slower over time as things were added
What this isn't
- A rebuild — if the foundation is the problem, optimisation is patching
- An SEO campaign — speed is one ranking input among many
- A guarantee of a green score — on some platforms, some scores aren't reachable
The three metrics
Largest Contentful Paint (LCP)
How long until the biggest visible element above the fold finishes rendering.
Target: under 2.5 seconds
Interaction to Next Paint (INP)
How quickly the page responds when someone clicks, taps or types, measured across all interactions.
Target: under 200ms
INP replaced First Input Delay in March 2024 — it's the more reliable metric.
Cumulative Layout Shift (CLS)
How much the page moves while it's loading. Ads, images without dimensions, fonts swapping in.
Target: under 0.1
What actually makes sites slow
Images
Almost always first, and usually the cheapest win. Uncompressed, wrong format, sized for desktop and shrunk in the browser, no lazy loading below the fold.
Too much JavaScript
Plugins and third-party scripts each add weight. On a mid-range phone, parsing that JavaScript costs more than downloading it.
Render-blocking resources
CSS and JS that must download and parse before anything appears. Fixable by inlining what's needed above the fold and deferring the rest.
No caching or CDN
Repeat visitors re-downloading what they already have; NZ visitors pulling every asset from a US server.
Layout shift
Images and ads without declared dimensions, so content jumps as things load. That's your CLS score, and it's usually a fifteen-minute fix.
Server response time
Sometimes the host, sometimes a slow database query behind the page. This is the one that needs actual investigation rather than a checklist.
How I measure
On a throttled mobile connection, on a real device profile — not a desktop Lighthouse run. Before and after, same conditions, so the numbers mean something.
Tools used: PageSpeed Insights for lab and field, Search Console's Core Web Vitals report for what Google sees, WebPageTest for waterfall charts, and CrUX history for up to 12 months of field data.
You get the before and after numbers either way, including when they're less impressive than either of us hoped.
What it costs
Speed audit and written report
From specified price. What's slow, why, what each fix would buy you, in priority order. Yours to hand to anyone.
Audit plus implementation
From specified price, depending on what the audit finds.
Ongoing monitoring
Part of a maintenance plan.
For context, NZ providers commonly bill ad-hoc developer work at NZ$80–150/hour.
Honest note:
On plenty of sites, the audit finds the main problem is images, and the fix is an afternoon you could do yourself. That happens often enough to be worth saying.
When you don't need me
Run PageSpeed Insights on your own site — it's free and it names the top issues. If it says "serve images in next-gen formats" and you can re-export your images, do that first and re-test. A lot of sites go from red to amber on images alone.
Come to me when the report is telling you things you can't act on, when the fixes need code, or when you've done the obvious and it's still slow.
Questions before we start
Related services
Speed optimization works best as part of broader development. You might also need a new site or app built.
Get a free speed check
Send me your URL and I'll tell you what's actually slow and roughly what it'd take to fix. No charge.
Get a free speed check