How do I speed up my online store?
Test a product page and a category page, not only the homepage, then fix the biggest cause first. In most stores that is an oversized main product image, too many apps and tracking scripts, or a slow server. Serve the main image at the size it is shown, load it first, delay chat and review widgets until the page is usable, then test again on a phone.
How to speed up your online store in an afternoon
Test the pages that sell
Paste the address of your best-selling product page into the speed test above. Run it again for your busiest category page. Shoppers decide on those templates, not on the homepage.
Read the phone verdict first
A page passes Core Web Vitals when LCP, INP and CLS are all good at the 75th percentile of real visits on phones. When a page has too few visits, the test falls back to real-user data for your whole store, then to the lab test.
Find the biggest delay
Open Where the time goes. Each line shows the seconds that fix could save, so start with the top line, not the easiest one.
Fix the main product image
Export it at the width it is shown on a phone, in WebP or AVIF, and give it
fetchpriority="high". Never lazy-load it; keep lazy loading for images further down the page.Cut apps and scripts
List every app, pixel, chat, review and popup script. Remove the ones nobody uses, and load the rest after the page can be tapped.
Reserve space for late items
Give galleries, banners, review stars and cookie bars a fixed space, so the Add to cart button does not jump while the page loads.
Check the server and compression
If the first byte takes more than 800 ms, fix hosting, caching or the database before you touch the theme, and turn on Brotli or gzip.
Test again, then let real users confirm
Run your next test on the full speed test page to confirm the lab result. Real-user data covers 28 days, so judge the fix a few weeks after you publish it.
The online store speed checklist
Work down this list on one product page, then roll each fix out to the template so every product gains.
- The main product image is exported at the size it shows on a phone, in WebP or AVIF
- The main image has
fetchpriority="high"and is never lazy-loaded - Related products, reviews and other images lower down are lazy-loaded
- Every app, pixel and widget earns its place; unused ones are uninstalled, not just hidden
- Chat, reviews and popups load after the page is usable
- Galleries, banners and review stars have space reserved, so nothing jumps
- Filters and Add to cart react to a tap within 200 milliseconds (INP)
- The server sends the first byte in under 800 ms, with page caching on
- Brotli or gzip compression is on for HTML, CSS and scripts
- Fonts are limited to the families and weights the design really uses
- The Arabic storefront is tested separately, with its own fonts and banners
- Cart and checkout pages pass the same checks as product pages
Where the time goes on a product page
Every LCP time splits into four parts, and each part has its own store fix.
| Part of LCP | What it is | Usual store fix |
|---|---|---|
| Time to first byte | The wait for the first byte of HTML | Better hosting, page caching, fewer database calls |
| Resource load delay | The gap before the main image starts to download | Put the image in the HTML, give it high priority, never lazy-load it |
| Resource load duration | The time to download the main image | A smaller file: the right size, WebP or AVIF, a CDN close to shoppers |
| Element render delay | The gap between the download and the image appearing | Fewer scripts and stylesheets blocking the page |
web.dev suggests keeping the two delays close to zero, so that most of the time goes on the first byte and the image download.
Who uses this test
Fashion and beauty stores
Large photo galleries are the usual reason the main image is slow. One test on a product page shows how many seconds the hero photo costs.
Electronics sellers
Spec tables, comparison widgets and review apps add scripts. The test shows which ones hold up the first tap on a filter.
Grocery and pharmacy delivery
Shoppers reorder in a hurry on their phones, so a category page that reacts slowly to a tap loses the basket.
Gift and flower shops
Demand peaks around Ramadan, Eid and Mother's Day, so test before each season rather than in the middle of it.
Social sellers opening their own site
Sellers who started on Instagram or WhatsApp test their first store pages before they spend on ads that send traffic there.
Store owners and their developers
Owners run the test and share the report link; developers work down Where the time goes in the order it gives.
Worked examples
A product page with a slow main image
Suppose your report shows an LCP of 4.2 seconds on phones, and Where the time goes puts properly sized images at the top. Export the main photo at the width it is shown on a phone, convert it to WebP, give it high priority and take lazy loading off it. The lab LCP drops on the next test; the real-user figure follows over the next 28 days.
A category page that feels stuck
Suppose LCP passes but INP is 380 milliseconds, in the needs work range. Scripts that run when a shopper taps a filter are the first suspect. Remove one tracking or popup app at a time, test after each, and keep the change that moves INP the most.
A new store with no real-user data
A store that opened last month may see "Not enough visitors yet for Google's real-user data." Use the lab result from a mid-range phone as your guide until traffic builds, and fix what it lists in order.
UAE tips for online sellers
Before the big sale week
White Friday and Black Friday next fall on Friday 26 November 2027, and most UAE retailers run the sale across the whole week. Finish speed work in the weeks before and add no new apps during sale week. The UAE holidays 2027 calendar lists the retail dates.
Ramadan evenings
Ramadan 2027 is expected to begin on Monday 8 February, subject to the moon sighting, and evenings get busy. Run a test in the evening too, when your server carries the most visitors.
The Arabic storefront
Test the Arabic version of a product page as well. It may load a different font and longer banners, and both can change the result.
WhatsApp and cash on delivery
Keep the WhatsApp order link and the cash on delivery note as plain text and links near the price, not inside a heavy widget. Make the link in the WhatsApp link generator.
Your own store and marketplaces
If you also sell on marketplaces, your own store competes on speed and trust. A fast product page with clear delivery times gives shoppers a reason to buy direct.
What changes in 2027 for store speed
Now
- Google judges Core Web Vitals on real visits at the 75th percentile, separately for phones and desktop, so your slowest quarter of phone visits sets the verdict.
- Google recommends good Core Web Vitals for success in Search and says they match what its core ranking systems reward, but it promises no ranking position.
- INP, which measures how fast a page reacts to taps and key presses across the whole visit, has replaced First Input Delay.
Next
- AdEngage UAE expects more shoppers in 2027 to land on a product page straight from an AI answer or a social video, ready to buy, so that page has to load fast and show price, delivery and returns at once.
- AdEngage UAE expects stores to keep adding scripts, from tracking to chat to shopping assistants, which makes a monthly speed test part of the routine rather than a one-off project.
- Test after every new app, theme update or campaign landing page, because each one can undo an earlier fix.
Speed test or full website check?
Use the speed test when you want to know why a page is slow and what to fix first. Use the website checker with Retail and ecommerce chosen when you want speed in context: it scores search, AI answers, trust and conversion as well, and ranks you against other UAE sites in retail. Both are free, and the full speed test page explains every metric in the report.