Your PageSpeed Score Isn’t the Whole Story

Google PageSpeed Insights has become one of the most recognised tools for assessing website performance. Whether you’re commissioning a new website, reviewing an existing one or simply curious about how your site compares, it’s often one of the first places people turn.

That isn’t surprising. The report is easy to run, the results are shown as a score out of 100 and the recommendations come directly from Google. A quick glance tells you whether your website sits comfortably in the green or whether there may be room for improvement.

For many organisations, that score quickly becomes the measure of success. The score is useful, but it doesn’t tell you everything.

PageSpeed Insights, powered by Google’s Lighthouse auditing tool, provides an excellent assessment of a website’s technical health. We use it throughout our own projects because it highlights genuine issues, helps us monitor performance over time and gives us confidence that the technical foundations of a website are sound.

Good Lighthouse scores matter. What they don’t do is measure every aspect of what makes a website perform well for the people using it.

A good Lighthouse score is worth aiming for

Sometimes discussions around PageSpeed become surprisingly polarised. One group argues that every website should score as close to 100 as possible, while another dismisses Lighthouse altogether, suggesting that user experience is all that really matters.

In reality, most websites sit somewhere in the middle.

Lighthouse exists for a reason. It measures aspects of a website that have a direct impact on how efficiently pages load and behave. Oversized images increase download times. Excessive JavaScript can delay interaction. Layout shifts create frustration as content moves around the screen. Accessibility issues affect how people, including those using assistive technologies, experience a page. Individually they may seem small, but together they influence how well a website performs for the people using it.

Many of Lighthouse’s recommendations improve more than technical performance. While Google doesn’t reward websites simply for achieving a high score, the factors Lighthouse measures often overlap with the signals Google uses when assessing search results. That’s one reason technical improvements are usually worth making. They benefit the people using your website first, while also supporting its long-term visibility in search.

When we develop or review websites, Lighthouse is one of our key tools. It gives us a consistent benchmark, highlights opportunities for improvement and often identifies problems that aren’t immediately obvious through visual inspection alone.

If a website consistently scores poorly, it’s usually worth understanding why. In our experience, once performance starts falling well below the green range, there is almost always something that deserves attention, whether that’s inefficient assets, accumulated technical complexity or broader issues with how the website has been built.

The score itself isn’t the problem. More often, it’s a symptom of something else.

What Lighthouse measures exceptionally well

One of the strengths of Lighthouse is that it focuses on things an automated system can measure objectively. It doesn’t have opinions. It simply analyses what happens as a page loads, how efficiently resources are delivered and whether recognised technical best practices have been followed.

That’s what makes Lighthouse such a useful tool. Like a health check, it gives you objective signs that something may need attention. When those signs begin to change, it’s usually worth taking a closer look to understand why.

It measures specific aspects of a website’s technical health rather than everything that influences its overall success.

A Lighthouse report can tell you whether images are unnecessarily large, whether JavaScript is delaying rendering or whether layout shifts are affecting stability. It can highlight accessibility issues that follow recognised standards and identify opportunities to improve technical efficiency.

What it can’t tell you is whether visitors found the information they needed, understood your content or completed the task they came to the website to achieve.

Those questions matter just as much to the success of a website, but they require human judgement rather than automated testing.

Is using AI always a risk?

Discussions about AI often become surprisingly black and white. Depending on who you listen to, it’s either the answer to every problem or something that should be avoided altogether.

Most technology doesn’t work like that, and AI is no different.

Organisations already rely on countless digital tools every day without thinking of them in terms of trust. Search engines help us find information. Spell checkers highlight mistakes before they’re published. Analytics platforms reveal patterns that would otherwise be difficult to spot. None of these tools replaces human decision-making. They simply help people work more effectively.

AI belongs alongside those tools.

Whether it’s appropriate depends on what you’re trying to achieve. In some situations it can save considerable time, help people organise information or explore ideas more efficiently. In others, it may add very little value at all.

The important question isn’t whether AI is involved. It’s whether it’s being used for a reason that genuinely improves the outcome.

Performance is about more than loading quickly

One of the most common misconceptions surrounding PageSpeed is that performance is simply a measure of time.

If a report says a page takes five seconds to load, it’s natural to imagine visitors waiting for the entire page to appear before they can begin using it. In reality, modern websites rarely behave like that.

What users experience is often far more important than the moment every last asset finishes downloading.

A well-engineered page may display its primary content almost immediately, allowing someone to begin reading or navigating while images further down the page continue loading in the background. Another page might appear complete before elements unexpectedly shift around the screen as late-loading assets arrive.

From a technical perspective, both pages may report similar loading times. From a user’s perspective, they feel completely different.

Google’s own performance metrics increasingly focus on how quickly meaningful content becomes available and whether pages remain stable while people are using them.

Performance isn’t only about how long a page takes to load. It’s about how confidently and comfortably people can begin using it.

Every improvement involves judgement

Every feature carries a cost. Images need to be downloaded, videos consume bandwidth and fonts add another resource for the browser to load. Analytics platforms, marketing tools and third-party integrations introduce additional requests before a page is fully rendered.

Individually, none of these things is necessarily a problem.

A promotional video may be one of the strongest ways for a university to showcase campus life. A carefully chosen typeface may strengthen brand recognition. Analytics and advertising platforms can provide valuable insight into how a website is performing.

It’s important to understand whether each feature provides enough value to justify the performance cost it introduces. The opposite is equally true.

This is where engineering judgement becomes important. Lighthouse identifies opportunities. Experience helps determine which improvements will have the greatest impact and where pursuing a perfect score may simply create unnecessary work for very little practical benefit.

There comes a point where optimisation stops being about improving the website and starts becoming about improving the number. Those two things are not always the same.

The score should start conversations, not finish them

Perhaps the most useful thing about Lighthouse is that it encourages questions.

Why has the score changed?
What is contributing most to the result?
Which recommendations are likely to make a meaningful difference?
Are there quick improvements available, or does the report highlight a deeper issue that deserves more careful attention?

Those questions are often far more valuable than the score itself.

We’ve seen websites improve significantly through relatively simple changes, such as properly optimising images (we can’t stress this one enough) or removing unnecessary assets. We’ve also seen websites where declining performance reflected years of individually reasonable decisions that had gradually accumulated over time. A new tracking platform here, another integration there, larger content uploads and additional functionality can all influence performance without any single change appearing particularly significant on its own.

That’s why performance shouldn’t be thought of as something you achieve once and never revisit. Websites evolve, and their technical health evolves with them.

Looking beyond the number

A strong Lighthouse score is something every organisation should value. It reflects good technical practice and provides reassurance that the foundations of a website are performing well. However, it’s also only one part of understanding how successfully a website performs.

Strong technical performance creates the right foundations. Long-term success comes from combining those foundations with clear content, accessible design and an experience that helps people achieve what they came to do.

Google’s Lighthouse report remains one of the best tools available for measuring technical performance, and it’s one we would encourage every organisation to use.

Numbers help us ask better questions. They rarely answer them on their own.