In this guide · 8 sections
- 01Field data vs lab score: what you are paying to improve
- 02What a website speed optimisation service should include
- 03Realistic expectations by platform
- 04Red flags when buying website speed optimisation
- 05How to check the work after it is done
- 06Keeping the site fast after optimisation
- 07Do it yourself or get help?
- 08Frequently asked questions
Short answer
A good speed service measures real-user field data (Core Web Vitals from Chrome users) before and after, fixes the main causes (images, scripts, hosting, caching), documents every change with a rollback plan, and checks again after 28 days. A green lab score alone proves little.
Field data vs lab score: what you are paying to improve
Google's Web Vitals guidance sets three Core Web Vitals: Largest Contentful Paint (loading) should happen within 2.5 seconds, Interaction to Next Paint (responsiveness) should be 200 milliseconds or less, and Cumulative Layout Shift (visual stability) 0.1 or less. Google recommends measuring them at the 75th percentile of page loads, on mobile and desktop separately.
PageSpeed Insights shows two different things on one page, and sales pitches often mix them up.
| Field data (CrUX) | Lab data (Lighthouse score) | |
|---|---|---|
| What it is | Real Chrome users' experience over the previous 28 days | One simulated page load on a test device |
| Used for | Whether real visitors have a good experience | Finding and debugging problems |
| Changes | Slowly, as 28 days of new visits replace old ones | Every run; Google notes scores vary from run to run |
| Pass mark | 75th percentile of LCP, INP and CLS all 'Good' | 90+ is shown as good |
Google's PageSpeed Insights documentation also notes that a good lab score does not necessarily mean real-user experiences are good, and that smaller sites may not have enough traffic for page-level field data, in which case it falls back to the whole site, or shows none.
What a website speed optimisation service should include
Whatever the price, a proper job has the same shape: measure, fix the largest causes in order, verify, and leave you with a record.
Speed optimisation deliverables
If you want to understand each fix, our Core Web Vitals fix guide explains the usual causes and the order to tackle them.
Realistic expectations by platform
How much an optimiser can change depends on how much of the site you control.
| Platform | What can usually be improved | Main limits |
|---|---|---|
| WordPress / WooCommerce | Hosting, caching, images, plugins, theme code, database | Page builders and many plugins add weight; some may need replacing |
| Shopify | Images, apps, theme code, third-party scripts | Hosting and core checkout are controlled by Shopify |
| Wix / Squarespace | Images, sections, embedded apps, fonts | Little control over code, hosting or caching |
| Custom-built sites | Almost everything | Depends on access to the code and the original developer's work |
On hosted platforms the biggest wins are often removing things: an unused app still loading scripts on every page, a chat widget, a large video in the hero section. A good optimiser will tell you which of these is costing the most, and let you decide.
Red flags when buying website speed optimisation
Speed is an easy service to oversell, because a lab score can be pushed up in ways that do nothing for customers.
A good offer
- Shows your current field data and explains what it means
- Promises improvement on named metrics, not a fixed score
- Backs up first and gives you a change log
- Tells you what the platform will not let them change
Red flags
- Guarantees '100/100 on PageSpeed'
- Only shows lab screenshots, never field data
- Wants to install a stack of extra optimisation plugins
- Hides or delays content to trick the lab test
Also ask who keeps admin access afterwards, and make sure any plugins or services added are in your name, not the agency's.
How to check the work after it is done
You can verify a speed job yourself with free tools.
Save the 'before' report
Run PageSpeed Insights on your top pages before work starts and save the results as PDFs or screenshots with the date.
Test the site yourself
After the work, click through key pages on your phone using mobile data. Check forms, checkout, menus and pop-ups still work.
Compare lab results the same day
Run PageSpeed Insights again on the same pages. Run each a few times, since lab results vary between runs.
Wait for field data
Field data covers the previous 28 days, so check again about four weeks later. That is the real result.
Check Search Console
The Core Web Vitals report in Google Search Console shows groups of URLs moving from Poor or Needs improvement to Good.
Keeping the site fast after optimisation
Most sites slow down again within months, because new content and tools are added without anyone checking speed: a 6 MB banner photo, a new booking widget, another marketing tag. Treat speed as part of maintenance.
- Resize and compress images before uploading; agree a maximum file size with whoever edits the site.
- Test speed after installing any new plugin, app or tracking script, and remove trial apps you did not keep.
- Check PageSpeed Insights and Search Console monthly, alongside updates and backups.
- Review third-party scripts quarterly: analytics, ads pixels, chat, heatmaps. Remove what nobody uses.
Our checklist for website maintenance services shows how speed monitoring fits with updates and security.
Do it yourself or get help?
If you are comfortable in your site's admin, image compression, removing unused apps or plugins and turning on caching are realistic DIY steps. Code changes, hosting moves and script deferral are where sites tend to break.
Our Speed & SEO Fix (S$390) covers the deliverables above for a typical small business site: baseline field and lab data, the main fixes, a change log and a follow-up check. For ongoing SEO beyond speed, see our SEO cost guide.
Frequently asked questions
No. The score is a lab simulation; Google's own documentation calls 90 and above good and notes that good lab data does not guarantee good real-user experiences. What matters is whether real visitors pass the Core Web Vitals thresholds in field data. Chasing 100 can mean removing useful features for little real gain.
PageSpeed Insights field data covers the previous 28 days of real Chrome users, so improvements appear gradually and fully after about four weeks. Lab results change immediately. The Core Web Vitals report in Search Console also updates over several weeks as new data replaces old.
Google's PageSpeed Insights documentation explains that lab scores vary between runs because of network conditions, the test machine's hardware and resource contention. Run each page a few times and look at the trend. Field data from real users is more stable because it is aggregated over 28 days.
Yes, within the platform's limits. Optimisers cannot change Shopify's or Wix's hosting or core code, but they can compress images, remove unused apps, simplify heavy sections and reduce third-party scripts. On these platforms, removing things often helps more than adding optimisation tools.
It can help. Google recommends good Core Web Vitals and says they align with what its ranking systems reward, but relevance and useful content matter more. The more immediate gain is usually fewer visitors giving up before the page loads, especially on mobile data.