Someone on your team ran your homepage through Google PageSpeed Insights, saw a score in the 40s or 50s, and now everyone's worried. Fair enough — it's a red flag worth taking seriously. But chasing the number itself is the wrong goal. The right goal is a site that loads quickly for real customers on real phones, and a 90+ pagespeed score is simply a handy sign you're getting there. Here's exactly what pushes that score up, in the order that gives you the best return for your effort.
What the PageSpeed score is actually measuring
Google PageSpeed Insights runs your page through Lighthouse and produces a single number from 0 to 100. That number is a weighted blend of several lab-measured metrics, not a direct measurement of "speed" in one dimension. The heaviest-weighted contributors include:
- Largest Contentful Paint (LCP) — how long the biggest visible element (usually a hero image or heading) takes to render.
- Total Blocking Time (TBT) — how long the main thread is jammed up by JavaScript, which is a lab stand-in for the real-world Interaction to Next Paint (INP) metric.
- Cumulative Layout Shift (CLS) — how much content jumps around as the page loads.
- Smaller contributors like First Contentful Paint and Speed Index.
If you want the full detail on what each of these metrics means and why Google cares, our Core Web Vitals guide walks through them properly. For this article, the short version is enough: the score is a proxy, built from a handful of real signals, run in a simulated lab environment rather than on your actual visitors' phones.
That matters because it's entirely possible to obsess over a lab score while your real customers have a perfectly fine experience — or the reverse. We've covered that gap in detail in how fast should a website load, which is worth reading alongside this one. Still, as a diagnostic target, 90+ is genuinely useful: it forces you to fix the things that also make the site faster for humans, not just for a scoring algorithm.
The checklist: what actually moves a pagespeed score
This is the order we work through with client sites, roughly from highest impact to lowest. Work down this list rather than jumping around — images and render-blocking code account for most of the damage on the average small business site.
- Compress and correctly size every image. Export images at the actual display size (not a 4000px camera photo squeezed into a 400px box) and compress them properly before upload.
- Serve modern image formats. WebP and AVIF are noticeably smaller than JPEG or PNG at equivalent quality, and every mainstream browser now supports them.
- Lazy-load offscreen images. Anything below the fold shouldn't load until the visitor scrolls near it.
- Set explicit width and height on images and embeds. This lets the browser reserve space before the image loads, which is the single biggest fix for layout shift.
- Turn on browser and server-side caching. Repeat visits and subsequent page loads should reuse cached assets instead of re-fetching everything. Our caching explained guide covers how this works in plain English.
- Put a CDN in front of your assets. A content delivery network serves images, CSS and JavaScript from a location close to the visitor instead of a single server, which shaves real time off every load. See CDN explained for how this fits together with caching.
- Minify CSS and JavaScript. Stripping whitespace, comments and unused code out of your files reduces what the browser has to download and parse.
- Defer or eliminate render-blocking resources. Anything that stops the browser painting the page — badly loaded stylesheets, synchronous scripts in the
<head>— should be deferred or loaded asynchronously. - Reduce third-party scripts to the ones you actually use. Every tracking pixel, chat widget and marketing tag is a small tax on load time. Audit them twice a year and cut what nobody looks at.
- Use decent hosting. No amount of code optimisation compensates for a server that takes a second and a half just to respond.
- Choose a lightweight theme and page builder (particularly relevant on WordPress, where bloated themes are a common culprit — see speed up your WordPress website).
- Trim your web fonts. Load only the weights and character sets you actually use — most sites load four or five font weights and only ever display two.
Quick wins vs. bigger jobs
Some of this you can do in an afternoon; some needs a developer.
| Action | Effort | Typical impact |
|---|---|---|
| Compress and resize images | Low | High |
| Enable a caching plugin/setting | Low | Medium–High |
| Add a CDN | Medium | Medium–High |
| Lazy-load images | Low | Medium |
| Set image/video dimensions | Low | Medium (CLS specifically) |
| Minify CSS/JS | Low (with a plugin) | Medium |
| Defer render-blocking scripts | Medium | High |
| Cut third-party scripts | Medium (needs a decision-maker) | Medium–High |
| Switch to better hosting | Medium–High | High |
| Rebuild on a lighter theme | High | High |
Images: where most sites lose the most points
If your pagespeed score is low, images are the first place to look. A photo straight out of a modern phone camera can run to several megabytes — completely unnecessary for a web page. Before anything gets uploaded to your site:
- Resize the image to the actual maximum width it'll display at (a full-width hero rarely needs to be wider than about 1920px).
- Compress it with a proper tool rather than eyeballing quality — most compression brings file size down substantially with no visible difference.
- Export in WebP or AVIF where your platform supports it, with a JPEG fallback if you need one for older browsers.
- Lazy-load anything below the fold so the browser isn't fetching your footer logo before it's painted the hero.
A Coffs Harbour dental clinic we worked with had a homepage sitting in the low 30s on PageSpeed Insights — almost entirely because every staff photo and treatment image was an unedited upload straight from a phone, several megabytes each, displayed at a fraction of their native size. Resizing and compressing the image library, switching the largest images to WebP, and moving from a budget shared host to proper managed WordPress hosting took the score into the low 90s without touching a line of code beyond what the hosting migration required.
Code and third-party scripts: the quieter culprits
Once images are sorted, the next biggest wins usually come from what's running in the background. Minifying CSS and JavaScript is close to a free win — most caching or optimisation plugins do it with a single toggle. Deferring render-blocking scripts takes a bit more care, because some scripts genuinely need to run early (analytics consent tools, for instance), but most don't need to block the page painting.
Third-party scripts deserve a proper audit, not just a technical fix. A live chat widget, a booking calendar, a review carousel, three different tracking pixels and an old marketing tool nobody uses anymore can each add meaningful weight. Ask, for each one: does this still earn its place? If a Ballarat cafe's booking widget only gets used twice a week, it might be worth linking to it instead of embedding it.
Why 100/100 usually isn't realistic — or the goal
A perfect 100 score sounds satisfying, but for most real businesses it's not achievable and not worth chasing. The moment you add a live chat widget, a booking system, a payment gateway, a review platform or an ad-tracking pixel, you're loading third-party code you don't control and can't fully optimise. Those tools do real jobs — taking bookings, answering questions, tracking ad spend — and ripping them out purely to chase a perfect score is a bad trade.
A genuinely fast site sitting in the high 80s or low 90s, with real customers experiencing quick load times, is a far better outcome than a 100/100 score on a stripped-back page with none of the tools your business actually needs. Treat 90+ as a strong target, not a finish line you must cross at any cost.
Key Takeaways
- The PageSpeed score is a weighted blend of lab metrics (LCP, blocking time, layout shift and more) — a useful proxy, not the whole story.
- Image optimisation delivers the biggest single improvement for most small business sites.
- Caching and a CDN, minified code, deferred scripts and decent hosting round out the checklist.
- Cut third-party scripts you don't genuinely use — each one adds weight.
- 100/100 usually isn't realistic once you add the booking, chat and tracking tools a real business needs, and it isn't necessary either.
- A high 80s to low 90s score with genuinely fast real-world loading beats a perfect lab score on a stripped-back page.
Frequently Asked Questions
What's considered a good PageSpeed score?
Anything from the high 80s up is generally considered strong, and 90+ is the commonly cited target for "good" performance. What matters more is whether real visitors experience a fast, stable page — the score is a useful proxy, not the end goal itself.
Does PageSpeed score directly affect Google rankings?
Page experience, including Core Web Vitals, is one of many ranking factors, but it's a relatively minor one compared to content quality and relevance. That said, a fast site improves conversions and reduces bounce rates regardless of any ranking effect, which makes it worth fixing either way.
Why is my mobile score so much lower than desktop?
Mobile testing simulates a slower processor and network connection, so the same page almost always scores lower on mobile than desktop. Because most Australian visitors browse on their phones, the mobile score is the one to prioritise — see our mobile speed optimisation guide for mobile-specific fixes.
Can a WordPress site realistically hit 90+?
Yes, plenty of WordPress sites score in the 90s, but it usually takes a lightweight theme, a good caching plugin, optimised images and reasonable hosting working together. A bloated page builder theme with dozens of unused plugins will struggle no matter what else you fix.
How much difference does hosting actually make?
A significant amount. Cheap shared hosting with slow server response times puts a ceiling on your score that no amount of front-end optimisation can lift. Moving to quality managed hosting is often the single change that unlocks everything else.
Should I fix everything myself or hire a developer?
Image compression, lazy-loading and basic caching plugin setup are reasonable for a business owner to handle. Deferring render-blocking scripts, auditing third-party code and hosting migrations are worth handing to a developer, since a wrong move can break site functionality.
Will removing plugins improve my score?
Often, yes — each active plugin can add its own CSS, JavaScript and database queries, even on pages where it isn't used. Running a full website speed audit will show you which ones are actually worth keeping.
How often should I re-check my PageSpeed score?
Check after any significant change — a new plugin, a design update, a new booking widget — and otherwise every few months as part of routine maintenance. It's also worth testing with more than one tool, since Google PageSpeed Insights, GTmetrix and WebPageTest each surface slightly different detail; our website speed testing tools guide compares them.
Get a website that scores well and performs well
A high pagespeed score is a means to an end: customers who don't bounce, forms that get filled in, and bookings that actually come through. If your site is scoring low and you're not sure which fix to tackle first, or you'd rather hand the whole job to someone who does this daily, have a chat with Pixel and Pine. We'll audit what's actually slowing your site down and fix it properly — not just chase a number.
