Quick answer

A useful speed review starts with the page your customer is trying to use and the point where it feels slow. A large opening image, a delayed menu and a form that jumps down the screen are different problems. Keeping the page address and the action beside the test results makes it easier to ask a developer for a specific repair.

For a UAE service business, we'd begin with the route to an enquiry: opening a service page on a phone, reading enough to decide, then calling or sending the form. The review below gives you a way to describe that route and compare proposed fixes.

On this page

Describe the delay before opening a report

Try the page as a new visitor would. If you normally stay logged in, use a private window too. Note whether the first useful content appears late, whether a tap waits for a response, or whether something moves just as you try to use it.

The page address matters. A fast homepage doesn't establish how a project gallery or booking page behaves. If your site has English and Arabic versions, try the actual pages your customers use in both languages. A different image or embedded tool can change the experience.

You can send a developer a short note such as: “On the mobile maintenance-service page, the text appears, then the enquiry button moves when the gallery loads.” A screen recording can make that easier to reproduce. It also gives you a useful check after the repair.

Read the two kinds of PageSpeed evidence

PageSpeed Insights can show real-user experience data alongside a controlled Lighthouse test. The two answer different questions. Real visits show the experience collected from eligible Chrome users; a lab test helps investigate a page under its test conditions.

Google says the real-user results cover a rolling 28-day period. A small or new site may have too little data for a page, or even its whole origin. In that case, “no data” doesn't tell you whether the page is fast or slow. Look at the label showing whether the report describes this page or the origin. Google's explanation of PageSpeed data

Keep the test date and mobile or desktop mode with the result. For a useful comparison, repeat the same test conditions rather than comparing a mobile result before the change with a desktop result afterwards.

Connect the metric to the customer's task

Google's Core Web Vitals describe loading, responsiveness and unexpected layout movement. The current “good” thresholds are assessed at the 75th percentile of visits, separately for mobile and desktop. Current Web Vitals definitions

Scroll the table to see all columns.

Current Core Web Vitals thresholds
MeasureGood thresholdA useful question for the repair brief
Largest Contentful Paint, LCP2.5 seconds or lessWhich main visible element is late, and what delays it?
Interaction to Next Paint, INP200 milliseconds or lessWhich interaction is slow, and what work runs when it happens?
Cumulative Layout Shift, CLS0.1 or lessWhich content moves unexpectedly, and can its space be reserved?

A developer still needs to investigate the cause. An image might need a different delivery size; a busy interaction might involve third-party code; a moving layout might need space reserved before media arrives. The score alone doesn't choose the fix.

Google uses Core Web Vitals in its ranking systems, but good results don't guarantee a top position. Your page still needs useful content and a clear mobile experience. Google's page-experience guidance

What two real mobile reports show

We tested two websites we've built on 16 November 2026, with three fresh-browser runs for each page. The screenshots show the middle Performance score from those three runs. We used Lighthouse 13.6.0 in local Chrome, with Moto G Power mobile emulation, simulated Slow 4G and 4× CPU slowdown. These are lab measurements of the tested pages.

Mobile Lighthouse lab comparison: Yasmin’s heading was the largest content element at 1.8 seconds, while Sherif’s hero image was the largest at 5.4 seconds.

Yasmin Teimoori: a strong result with more to inspect

The Yasmin Teimoori project includes a real-estate page at yasminteimoori.com/real-estate. Its three Performance scores were 93, 99 and 98, giving a median of 98. In the report shown, Largest Contentful Paint was 1.8 seconds, Total Blocking Time was 110 milliseconds and Cumulative Layout Shift was 0.

Mobile Lighthouse report for Yasmin Teimoori, 16 November 2026: performance 98, LCP 1.8 seconds, TBT 110 milliseconds, CLS 0.

Lighthouse identified the opening heading as this page's LCP element. That detail matters when deciding what to investigate. The report also flagged image-delivery opportunities for property photographs, so there's useful work to assess even with a high overall score.

Sherif EMW: follow the element that appears late

For the Sherif EMW project, we tested sherifemw.ae. Its three Performance scores were 74, 78 and 79, giving a median of 78. The middle report recorded First Contentful Paint at 1.8 seconds and Largest Contentful Paint at 5.4 seconds, with Total Blocking Time of 110 milliseconds and Cumulative Layout Shift of 0.

Mobile Lighthouse report for Sherif EMW, 16 November 2026: performance 78, LCP 5.4 seconds, TBT 110 milliseconds, CLS 0.

Here, Lighthouse identified the opening plant-room photograph as the LCP element. Some content appeared earlier, while that large visible element took longer under the simulated mobile conditions. We'd trace how the hero image is fetched and displayed before choosing a repair, then repeat the same page test to assess the change. The layout-shift result helps keep that loading investigation separate from content moving around.

The report also flagged a four-hour cache lifetime for images. Longer browser caching can help repeat visits. This particular audit is unscored, so the warning itself does not subtract Performance points. The effect of any caching change would need its own test, with a clear distinction between a new visitor and a returning one.

The range across the runs is worth keeping with the screenshots. Google explains why Lighthouse scores vary, and these reports don't establish real-user Core Web Vitals or an increase in enquiries. Total Blocking Time is a lab metric; it isn't an INP measurement.

Use a worked repair brief

Here's a fictional maintenance business reviewing its service page. These are example observations, rather than test results from an NST client.

Scroll the table to see all columns.

Why is your business website slow on mobile? — worked worksheet
ObservationWork to investigateAcceptance check
The opening photo arrives well after the headingIdentify the image request, delivered dimensions and loading priorityRepeat the same mobile test and compare when the main content appears
The enquiry button moves when photos loadCheck whether media has reserved dimensionsReplay the page load and confirm the button stays in place
The menu pauses after a tapRecord the interaction and inspect the work it triggersRepeat the interaction on an agreed phone and check the response
The contact form shows success but the office cannot find the messageTrace the form through delivery and the receiving inboxSend an identifiable test enquiry and confirm staff receive it

That last row belongs in the review even though it isn't a speed score. A faster page still needs to deliver the enquiry. It helps to keep the technical measurement and the customer's completed task in the same brief.

Agree on the repair and the retest

Before commissioning the work, ask which pages and shared templates it affects. A gallery change may help several project pages; removing a booking script could break a useful function. Knowing those dependencies lets you agree what needs checking before release.

The proposed scope should identify the suspected cause, the change and how the team will compare the result. A screenshot of a higher score is easier to assess when it comes with the page address and test conditions. Changes to field data take time to appear in a rolling reporting window.

It's also useful to know what the work leaves untouched. If the supplier is fixing image delivery, a slow external booking provider may remain a separate item. Our website quote guide gives you a worksheet for putting these responsibilities into the same request.

Discuss the pages that need attention

We can help scope the website work around the delay you can reproduce and the enquiry route you need to preserve. See our Website Development service for the broader build and support discussion.

Send NST the page address and the problem you see, along with the phone or browser you used and any recent change to the site.