An email lands in your inbox from an agency you have never heard of. It includes a screenshot of your PageSpeed Insights result, a red or orange circle on mobile, and a promise: they can get you to 100. Or a competitor mentions that their site scores 98. Or you simply keep looking at the number and wondering whether anything below a perfect score is failing.
So, is a 100 PageSpeed score necessary? For almost every small business website, no. A higher score is often a sign of a faster page, and moving from poor to decent is usually worth real effort. But the last stretch towards 100 tends to cost far more than it returns, and sometimes it is reached by tricks that make the site worse for the people using it. This post is about where to stop, and how to tell when a score is good enough.
What the number actually rewards
The Performance score is a weighted summary of a few lab measurements taken during one simulated page load: how soon something appears, how soon the main content appears, how long scripts block the page, how much the layout moves and how quickly the screen fills in. It is calculated on an emulated phone with a deliberately slow connection, from Google’s servers, with an empty cache.
That is a fair and useful stress test. It is not the same thing as your customer’s experience. The score knows nothing about whether your menu is easy to use, whether the product photos are good enough to sell from, whether the enquiry form works, or whether your visitors arrive on fast home broadband or crowded mobile data. It also does not measure what happens after the first load: clicking through to a second page, which on a well-cached site is often much faster.
Is a 100 PageSpeed score necessary for customers?
Customers do not see your score. They see a page that appears quickly or slowly, buttons that respond or lag, and content that stays still or jumps around. Those experiences line up with the Core Web Vitals, and Google publishes clear thresholds for what “good” means for each, measured on real visits rather than a lab run. A page can pass all three in the field with a lab score well below 100.
Where the score helps is as an early warning. A mobile score deep in the red almost always means something real is wrong: huge images, a blocking script, a slow server. Fixing those helps customers. A score of 88 against one of 97, on the other hand, rarely corresponds to a difference a person would notice on their phone.
What getting the last points usually costs
On a typical WordPress or WooCommerce site, the final points are usually held back by things the business chose on purpose. Getting rid of them is possible. Whether you want to is another matter.
| Often holding the score back | What removing it costs you |
|---|---|
| Analytics and advertising tags | Knowing where enquiries come from, and measuring ad spend |
| Live chat or a WhatsApp widget | A contact route some customers prefer |
| Embedded maps, video and reviews | Trust signals and information visitors look for |
| Brand web fonts | Part of how the site looks and feels |
| A page builder the team knows how to edit | Being able to update pages without a developer |
| A large, high-quality hero photo | The first impression, especially for visual businesses |
Many of these can be made lighter without being removed: delaying a chat widget until someone scrolls or taps, replacing a video embed with a preview image that loads the player on click, trimming font weights, sizing the hero properly. That middle ground is where most of the sensible work happens. Pushing past it, to remove the thing altogether purely for the score, is where you start trading real business value for a number.
Diminishing returns, in plain terms
The early fixes on a slow site are often cheap and large: compressing images, turning on page caching, removing a plugin nobody uses. Each one can move both the score and the real experience noticeably.
Later fixes get harder and smaller. Generating critical CSS, splitting scripts per page, removing a theme’s built-in features, rewriting a section without the page builder. Each takes more time, carries more risk of breaking something, and needs ongoing care when the site changes. The score keeps creeping up, but the person on the phone gains less and less from each step.
The cost does not end on the day the work is done, either. A site tuned to the edge is fragile. The next plugin update, a new marketing tag, or one large image uploaded by a member of staff will knock it back down, and someone has to notice and fix it again.
Be wary of tricks that inflate the score
Because the score is a single number, some people optimise for the number instead of the visitor. Things we would treat as red flags if a provider proposed them:
- Detecting the testing tool and serving it a stripped-down page. The test looks great; real visitors get the full, slow page. This is misleading, and showing different content to Google’s tools than to people is not something you want associated with your site.
- Delaying every script until the first interaction, including ones the page needs. Done carefully, delay is useful. Done blindly, menus, forms and sliders do nothing on the first tap.
- Hiding the main content until the visitor scrolls or taps, so the test measures a smaller, emptier first screen than the one people actually wait for.
- Promising a specific score as the outcome. Nobody can promise one, because the score varies between runs and depends on things like third-party scripts and Google’s own changes to the tool.
If you see a quote built around a fixed score such as “90+ on mobile” or “we will get you to 100”, ask what they will remove or change to get there, and what happens to your analytics, chat and forms.
When a higher score is worth chasing
There are cases where pushing further makes sense:
- Paid ad landing pages, where every visitor costs money and many arrive on mobile. A lean, dedicated page is often worth tuning harder than the rest of the site.
- Sites whose visitors are mostly on slow connections or older phones, where the lab’s harsh conditions are close to reality.
- Stores with many products and heavy pages, where gains on a category or product template apply across hundreds of pages at once.
- Sites failing Core Web Vitals in the field, where the lab score and the real-user data both point at the same problem.
In each case, though, the target is the experience and the field metrics. The score just tells you whether you are heading the right way. On the separate question of rankings, our post on whether Core Web Vitals affect SEO sets out what Google has actually said.
How to decide your score is good enough
A practical checklist. If most of these are true, your effort is probably better spent on content, offers or the enquiry process than on more points:
- Where field data exists, the page passes the Core Web Vitals assessment on mobile.
- In the lab, LCP, Total Blocking Time and CLS are in or near the green, even if the overall score is not.
- The page looks usable quickly on a mid-range phone on mobile data, and the menu and buttons respond promptly.
- The remaining suggestions mostly concern third-party tools you have decided to keep.
- The results are reasonably steady across several test runs, rather than occasionally good.
- Your most important pages (landing pages, product pages, checkout or contact) are as healthy as the home page.
The metrics themselves, and what “good” means for each, are explained in Core Web Vitals in plain language. To check results fairly, test the same way each time, as described in our guide on how to test a website’s speed properly. And for how scores, vitals and fixes fit together across the whole site, see the website speed and performance guide. Good enough, measured honestly, beats perfect on a screenshot.
Frequently asked questions
What PageSpeed score should a small business website aim for?
There is no single right number. A reasonable aim is to get the mobile score out of the red, keep the main lab metrics close to Google’s good thresholds, and pass Core Web Vitals in the field where your site has data. Beyond that, extra points usually cost more effort than they return for visitors.
My competitor scores higher than me. Does that matter?
Only if their site is noticeably quicker for real people. Compare how both sites feel on the same phone, and look at field data where it exists. A higher lab score can come from a simpler page with fewer features, and copying that simplicity is only sensible if those features are not helping your customers.
Can I get a perfect score with WordPress?
It is possible on a simple WordPress site with a lightweight theme, good hosting, careful caching and very few outside scripts. It becomes much harder once you add analytics, chat, embeds, a page builder or WooCommerce. Getting close is often realistic; holding a perfect score as the site changes is the harder part.
Will a perfect score push my site higher in Google?
Do not expect it to. Google has described page experience as one consideration among many, with relevant, useful content coming first, and where it looks at speed it uses real-visitor data rather than a single lab score. Moving from a very slow page to a reasonably fast one can help visitors; the last few points are unlikely to change rankings.
Getting a slow page to a comfortable speed is the worthwhile part, and it is what our website speed optimisation service focuses on, without promising you a particular number.
Next in this series: GTmetrix Waterfall Chart Explained, Row by Row
