PageSpeed Isn’t What You Think: Real Data from 50 Sites

I love PageSpeed. It’s one of my favourite benchmarks one of the few parts of web performance that feels measurable and honest. You can’t spin it with marketing or hide it behind a shiny homepage. A site is either fast or it isn’t.

I’ve spent years chasing milliseconds, stripping out unnecessary scripts, and tuning servers just to hit that green score. There’s real craft in performance optimisation, and when it’s done well, it makes a site feel effortless.

But this all started with a chat between me and two devs I really admire both working for huge fashion brands. They were showing off their new sites, talking about design, UX, integrations, campaigns all the usual stuff you’d expect from big launches. But one thing never came up: speed.

Not once.

They didn’t mention performance scores, load times, or Core Web Vitals. It was like speed didn’t matter. And that hit me.

I build brands I don’t just maintain them and for me, performance has always been part of the brand. A fast site feels like a sharp brand. It says the business is on it, that they care about the details. So hearing two developers at that level completely dismiss PageSpeed made me question my own priorities.

Maybe I’d been treating PageSpeed like the benchmark for success when, in reality, it’s just one piece of the puzzle. Maybe I’d been chasing a metric that means more to executive teams than it does to the people who actually use the site.

That conversation was the push. I wanted to see what the data really said whether PageSpeed actually correlates with success, or if I’d been giving it more credit than it deserves.

So I stopped guessing and measured it.

---

What I Set Out to Do

The goal was simple: separate belief from reality. I wanted to see how PageSpeed scores align with real-world rankings, not test environments or best-practice guides.

I chose ten everyday search terms that represent real small businesses florists, furniture stores, kitchen fitters, electricians, and more. For each term, I pulled the top five organic results and tested each page using Google’s PageSpeed Insights API (mobile).

That gave me a dataset of fifty live pages covering trades, retailers, and service providers. I measured performance, Core Web Vitals, and Lighthouse category scores. Every test ran under the same conditions within a 24-hour window to keep things consistent.

From there, I looked for patterns:

• Do faster sites actually rank higher?
• Which industries perform worst?
• And which metrics cause the biggest drag?

---

Findings and Analysis

I didn’t expect every site to be perfect but the results still surprised me.
Most of the sites I tested weren’t fast. In fact, the majority sat comfortably below Google’s “good” threshold. And yet, many of those same sites ranked first or second for competitive keywords.

The more I looked at the data, the clearer the pattern became: speed helps, but it’s not the deciding factor.

Rank	Average Performance Score
1	74.8
2	60.4
3	53.3
4	62.6
5	58.8

---

Average Performance by Industry

Across all fifty pages, the average PageSpeed score landed at 62, with most industries sitting between 50 and 70.

Retail sites tended to be slower, weighed down by large imagery, third-party scripts, and tag managers. Trades and service sites typically simpler builds performed slightly better but still far from Google’s ideal.

What stood out most is how consistent those results were. None of these industries, even the best performers, were close to perfect.

Average PageSpeed scores across small-business industries ranged from 49 (fashion) to 74 (jewellery).

The data confirms what most developers already know: real-world websites are messy. Small-business sites don’t have the infrastructure or budgets of enterprise setups, and even high-performing brands struggle to stay “green.”

---

Does PageSpeed Correlate with Rankings?

To test the core question whether faster pages rank higher I compared each page’s performance score against its Google position. The correlation was -0.3, which is statistically weak.

In plain terms: there’s no reliable connection between speed and ranking.

Some of the slowest sites ranked first. Speed alone didn’t predict visibility.

This aligns with what I’ve seen in practice. Google values usability and relevance but speed is only one small piece of that puzzle. A site that loads in three seconds and clearly answers the search intent will always beat a lightning-fast page that doesn’t.

---

Core Web Vitals in the Real World

When looking deeper into Core Web Vitals, most sites passed at least one metric, but very few passed all three.

Largest Contentful Paint (LCP) the time it takes for the main content to appear was consistently the biggest problem. Layout stability (CLS) was generally solid, suggesting most modern templates and builders are handling that well.

Keyword	LCP (s)	CLS	TBT (ms)
bathroom fitter	4.1	0.06	130
corner sofa	4.4	0.06	150
electrician	3.7	0.08	115
engagement rings	3.2	0.02	98
garden furniture	3.8	0.01	140
kitchen renovations	4.5	0.03	165
kitchen worktops	5.2	0.25	182
summer dresses	4.3	0.03	149
wedding photographer	3.9	0.03	135

LCP failures are widespread across all industries, while CLS and TBT remain within acceptable limits.

What’s clear is that the Core Web Vitals “ideal” is still out of reach for most small-business websites. That doesn’t mean they’re bad experiences it means they’re realistic ones.

In short: most small-business sites aren’t slow they’re just not perfect. LCP is the one metric that consistently drags them down, and in my experience, it almost always comes back to images. Oversized hero banners, uncompressed product photos, and lazy-loading content that should appear immediately are the biggest culprits.

I’ve yet to see a “slow” site that wasn’t really just a site full of heavy, unoptimised images. The fix isn’t complicated it just needs a little attention to what loads first.

---

Quality Beyond Speed

While Performance averaged in the low 60s, the other categories consistently landed in the 80s and 90s across every keyword. That gap is the real story even when a site isn’t technically fast, it can still deliver strong SEO structure, clean markup, and accessible design.

The reason’s simple: these scores measure stability and clarity, not speed. Tools like Lighthouse reward semantic HTML, proper headings, descriptive alt text, and security best practices all things most CMS platforms handle well out of the box.

Keyword	Performance	SEO	Accessibility	Best Practices
bathroom fitter	61	91	86	93
corner sofa	58	88	82	89
electrician	63	92	87	94
engagement rings	66	90	88	92
garden furniture	59	86	83	90
kitchen renovations	62	89	85	91
kitchen worktops	54	87	84	88
summer dresses	57	92	86	93
wedding photographer	64	90	89	92

Even with modest performance, most sites scored above 85 in SEO and Accessibility. That suggests Google values overall usability and technical hygiene more than shaving off a second of load time.

Even with modest performance, most sites scored above 85 in SEO and Accessibility.

For small-business sites, this is good news. You don’t have to chase perfect PageSpeed scores focus on accessibility, structured content, and best practices. They’ll earn you more trust, better rankings, and a smoother user experience than any 100/100 speed score ever will.

1. Google’s Core Ranking Signals

PageSpeed (performance) is only one small ranking factor among hundreds.
Core Web Vitals matter, but their impact is marginal compared to:

• Content relevance
• Authority (links and trust signals)
• UX signals like engagement and clarity

So improving speed from a score of 70 → 90 might make your site feel better, but it rarely moves the SEO needle unless your site is truly slow to begin with.

2. Accessibility and Best Practices Build Trust

Google uses Lighthouse’s Accessibility and Best Practices categories as proxies for overall quality not direct ranking factors, but strong indicators of:

• Usability for real visitors (which influences bounce and conversion)
• Technical stability (no security or structural issues)
• Readable markup and proper semantics (which help crawlers interpret your pages)

Those translate indirectly into better rankings because they create a site that’s easier for both users and search engines to navigate and trust.

3. Small Businesses Win by Balancing Effort

Most small-business sites are resource-constrained.
Spending 20 hours to improve a PageSpeed score from 85 to 95 produces diminishing returns, while improving content clarity, accessibility, or schema markup delivers measurable benefits for SEO and UX.

---

The Outliers That Prove the Point

The most interesting results came from the outliers big retail brands ranking high despite poor performance scores.

Next.co.uk ranked first for “summer dresses” with a PageSpeed score of 49. QuizClothing.co.uk ranked second with 43. SCS.co.uk ranked second for “corner sofa” with 48.

These aren’t anomalies. They’re proof that PageSpeed is just one input in a much larger equation. Brands like these win because Google doesn’t see them as pages it sees them as entities. They’ve built years of backlinks, brand searches, and behavioural trust. Even with slower sites, users click, stay, and convert, and Google rewards that engagement loop.

In short, authority protects inefficiency. Big brands can afford technical debt because they’ve earned search confidence.

Pro Tip: Smaller sites can’t compete on brand equity, but they can win on precision. Target specific long-tail intent, keep your content structure tight, and make every click frictionless. Google’s not looking for the fastest site it’s looking for the best answer delivered fast enough.

---

Interpretation & Takeaways

PageSpeed has always been one of my personal benchmarks for success in development. I’ve chased green scores for years, tuning every line and millisecond to squeeze out that perfect result. It’s been a metric of pride proof that a build was clean, efficient, and technically sound.

But this data forced me to take a step back. While I haven’t tested every site across these keywords, the results strongly suggest that speed isn’t the deciding factor. It supports rankings, but it doesn’t define them. Relevance, usability, and trust still carry more weight.

That balance is the heart of Google’s current E-E-A-T model. PageSpeed supports the experience side of that equation, but it’s not enough on its own which I explored in detail in my State of SEO 2025

Most real-world websites live in the middle fast enough to deliver a good experience, but rarely perfect. And that’s fine. A stable, intuitive site that serves its purpose will always beat one that’s optimised into fragility.

That said, PageSpeed will always be a factor in how I build and review websites. It might not be the be-all and end-all, but as a developer, hitting best practices and performance benchmarks will always matter. They represent craft, care, and respect for the user.

Perfection isn’t the goal consistency and experience are. PageSpeed is still part of that pursuit, just not the whole picture.

Pro Tip: Treat PageSpeed as part of your conversion stack, not your ranking strategy. Measure how speed changes bounce rate, time on page, or conversion rate those are the metrics that actually prove performance is working.

---

Authority Still Rules

The outlier brands make this painfully clear.
Big names like Next, Quiz, and SCS still rank at the top with PageSpeed scores in the 40s because they’ve earned something far harder to replicate authority. Years of backlinks, brand searches, and user trust carry more weight than a few extra milliseconds of load time ever will.

For smaller businesses, that’s frustrating but it’s also the roadmap. Speed won’t fix weak authority. You can’t optimise your way past thin content or a lack of reputation.

For small and mid-sized sites, the smarter move is balance.
Build authority through content that earns links and credibility. Keep the site technically sound and easy to use. Optimise the things that actually move the needle not the ones that look good in an audit.

Pro Tip: PageSpeed can open the door, but authority keeps you in the room. Focus on building relevance and reputation first; performance should support that story, not try to replace it.

---

What “Good Enough” Really Means

The takeaway isn’t to ignore PageSpeed it’s to understand where it actually matters.

A site that loads in under three seconds on mobile, stays visually stable, and delivers what users came for is already doing its job. If that site scores 68 instead of 92, it’s not broken it’s realistic.

And that last part delivers what users came for is what really matters now. Performance creates trust, but helpfulness creates retention. I broke that down further in Why ‘Helpful’ Isn’t a Buzzword It’s the New Homepage

“Good enough” doesn’t mean careless. It means knowing when to stop chasing technical perfection and start prioritizing what users and search engines actually reward. A fast, reliable site is the baseline not the endgame.

Here’s what “good enough” really looks like in practice:

• Sub-3-second load time on mobile. Users won’t notice or care beyond that.
• All Core Web Vitals in the green or close. A passing grade is good enough chasing perfect scores rarely delivers measurable benefits.
• Stable layout and responsive interaction. CLS and TBT matter because they impact usability, not ranking.
• Consistent performance under real-world conditions. A 75 score in Lighthouse that feels fast on 4G is better than a 95 that collapses under tracking scripts in production.

Perfection is expensive. Every hour you spend chasing another five points is an hour you’re not improving content, UX, or credibility the things that actually grow traffic and conversion.

As a developer, I’ll always care about performance, but I’ve learned that balance beats obsession. The goal is a site that feels fast, works consistently, and earns trust. That’s what users notice. That’s what Google notices too.

Pro Tip: Treat PageSpeed like tire pressure check it regularly, keep it healthy, but don’t rebuild the car to hit a perfect PSI. Once your site loads quickly, focus on clarity, structure, and user flow. That’s where the real wins are.

---

The Real Benchmark

After analysing 50 top-ranking pages across multiple industries, one conclusion stands out:

PageSpeed matters, but perfection doesn’t.

Performance should always be measured, maintained, and respected, but it’s not the deciding factor it’s often made out to be. The data made that clear: scores ranged from the 40s to the 90s, yet rankings barely shifted. The web doesn’t reward precision it rewards relevance, clarity, and trust.

If your site loads in a few seconds, feels stable, and gives users exactly what they came for, you’re already ahead of most of the web. The rest is polish.

Somewhere along the way, PageSpeed stopped being a diagnostic signal and became dogma. Every audit, plugin, and agency pitch has turned “fix your performance” into a commandment. But the truth is simpler speed helps users; it doesn’t decide visibility.

That doesn’t mean you should ignore it. It means you should treat it with context. Measure performance because it affects experience, not because you expect it to move rankings on its own.

For small businesses, the real benchmark isn’t a perfect score it’s a site that:

• Converts visitors into customers without friction.
• Feels fast on mobile and desktop.
• Stays visually stable and predictable.
• Delivers useful, trustworthy content.

PageSpeed matters but only up to the point where it stops helping.

Once that threshold is met, chasing perfection is wasted effort. The web rewards relevance, authority, and usability far more than decimals on a dashboard.

---

Final Thought

If you build or manage websites, focus on what users actually feel. The metrics, scores, and reports are just signals useful, but never the full story. What matters is the moment when someone lands on a page and it simply works: it loads, it’s clear, and it earns their trust.

Optimise until the experience feels smooth, then stop chasing decimals. Put that energy into what performance can’t give you stronger content, clearer structure, and real connection.

Because in the end, users don’t remember how fast your site scored. They remember how fast it made sense.

---