We just ran greenleafcreative.co.uk through Google’s PageSpeed Insights, and the results are the kind you actually want to show off: 91 for Performance, and a clean sweep of 100s for Accessibility, Best Practices, and SEO. It’s a good excuse to talk about why this tool matters more than most business owners realise — and where the line sits between chasing a perfect score and actually building a site people want to use.
What PageSpeed Insights actually measures
PageSpeed Insights scores a page across four areas — Performance, Accessibility, Best Practices, and SEO — using real Lighthouse audits, and increasingly reflects what Google calls Core Web Vitals: loading speed, interactivity, and visual stability. These aren’t arbitrary numbers. Google has confirmed page experience signals feed into search ranking, particularly on mobile, where competition for the top spot is fiercest.
A local example: “website designers in Perth”
Here’s where it gets practical. Say two agencies are both trying to rank for “website designers in Perth.” Agency A has a slick, animation-heavy homepage that scores 54 on mobile Performance. Agency B has a clean, fast build scoring 91, like ours. All else being roughly equal — similar content depth, similar backlinks — Google’s page experience signals nudge the ranking toward Agency B. But it’s not just rankings: a Perth-based buyer searching on their phone in a coffee shop is going to bounce off a 4-second load far faster than they’ll bounce off a plain-but-instant one. For a local service business, that’s the difference between a lead and a lost tab. Ironically, the agency with the worse score is often the one trying hardest to look impressive — which is exactly the tension worth talking about.
So when is it worth sacrificing score for a feature?
This is the real question underneath all of this, and the honest answer is: it depends on what the feature is actually buying you.
A few working rules we use:
- Protect the 90+ “green” band on mobile if you can. Below 90, you’re still fine — 50-89 is Google’s “needs improvement” zone, not a penalty box — but once you’re below 50, you’re in territory where the ranking signal and the human bounce rate both start working against you. A hero video that drags Performance from 91 to 85 is a reasonable trade. One that drags it to 40 usually isn’t.
- Ask what the feature is replacing. A background video that replaces a static hero image is rarely worth 20+ points of Performance for most local service sites — the conversion lift is marginal and the cost is real. A product configurator or interactive quote tool, on the other hand, might justify a real hit because it directly drives revenue.
- Mitigate before you sacrifice. Lazy-load the video, defer non-critical animation JS until after first paint, serve a compressed poster image on mobile instead of autoplaying video at all. Often you can keep 80% of the visual impact for 20% of the performance cost — which is a much easier trade to justify.
- Test on mobile, not desktop. Desktop connections and CPUs hide a multitude of sins. If a feature only tanks the score on desktop, it’s usually fine. If it tanks mobile, that’s where most of your local search traffic actually lives.
The takeaway
A 91/100/100/100 result isn’t just something to screenshot — it’s evidence the groundwork for good SEO is already there, and it’s a baseline worth protecting. Not every feature is worth defending that number for. But when you do add something that costs you points, do it with your eyes open: know whether you’re trading it for genuine conversion value, or just for something that looks good in a pitch deck.