Performance
How I got Wax London's homepage on screen sooner
On a phone, the hero went from 3.0s to 2.2s, which meets Google's good mark. On a computer it went from 3.2s to 2.8s. The header jump on a phone went to zero.
- WebPageTest
- Shopify
- Liquid
- JavaScript
Problem The first screen was late
I worked with Wax London on a homepage that had one job before anything else: show the hero. If that photo is late, or the top of the page jumps, people wait, then the brand feels slow.
That is a common first-screen problem on fashion homepages. The hero is the largest paint. Two announcement bars with different heights make the header drop. Third-party scripts for analytics, consent, and tests then compete with that first screen.
I measured the homepage in WebPageTest, on a phone and on a computer, then shipped the changes that would move Largest Contentful Paint and Cumulative Layout Shift. The live page is waxlondon.com.
If the hero lands sooner and the announcement bar keeps a steady height, the first screen should feel stable and ready, instead of late and twitchy.
On a phone, Largest Contentful Paint went from 3.0s to 2.2s. That is 26.7% faster, and it meets Google’s good threshold of 2.5s. The hero is now on screen by three seconds. That frame was missing in the before filmstrip.
On a computer, Largest Contentful Paint went from 3.2s to 2.8s. That is 12.5% faster. It is closer to 2.5s. It is not there yet.
These figures are lab runs, not Chrome User Experience Report field data.
Work What I changed
This was not a theme rewrite. I went after the first screen.
Got the hero on screen sooner
Largest Contentful Paint moved on both devices. On a phone the hero is now visible by three seconds, which the old filmstrip never showed.
Matched the announcement bars
The theme bar and the global bar had different heights. When one replaced the other, the page jumped. Same height on both, and the jump stopped.
Left desktop shift alone
Desktop Cumulative Layout Shift was already 0.002. CookieScript still adds a little. I did not chase a number that was already inside the budget.
Checked the leftover script cost
A blocked-script run showed Triple Whale, CookieScript, and Convert still cost Speed Index and Largest Contentful Paint. That is leftover work, not a live change.
Results What moved
WebPageTest on the homepage. Phone and computer. The phone now meets the Largest Contentful Paint bar. Desktop is closer. Layout shift on a phone is gone.
Phone
2.2s
Largest paint
from 3.0s
Computer
2.8s
Largest paint
from 3.2s
Phone
0
Layout shift
from 0.003
Computer
0.002
Layout shift
already inside budget
Largest paint
Largest Contentful Paint is the moment the main content shows up. On this homepage that is the hero. Mobile now meets Google's good mark. Desktop is still 0.3s over 2.5s.
Mobile after is 27 March to 28 March 2025. Lab data, not field data.
Phone · Largest paint
2.2s on a phone
Before, the first meaningful hero frame sat at 3.0s. After, it sits at 2.2s. That is inside Google's 2.5s good mark. People see the jacket, not a blank phone screen.
Computer · Largest paint
2.8s on a computer
The same hero moved from 3.2s to 2.8s. That is better. It is still 0.3s over the 2.5s target. Desktop still needs image and script work to close that gap.
Layout shift
The theme announcement bar and the global announcement bar were not the same height. When the global bar took over, the page dropped. Matching the heights removed the mobile jump. Desktop shift was already 0.002, well inside the 0.1 budget.
CookieScript still adds a little desktop shift. I did not spend the next week shaving 0.002.
Phone · Layout shift
The jump is gone
Same height on the theme bar and the global bar. The first screen stays put. People can tap a link without the header sliding out from under their thumb.
Takeaway What this means
This is the kind of homepage work I do for Shopify stores. I get the hero on screen, keep the header still, then measure what third-party scripts still cost.
Triple Whale, CookieScript, and Convert still sit on this homepage. I blocked them in a WebPageTest diagnostic, not on the live site. Speed Index dropped from 6.603s to 2.838s. Largest Contentful Paint dropped from 3.129s to 2.375s. That leftover cost is why a store can win the hero and still feel unfinished.
The domains were api.config-security.com, whale.camera, geo.cookie-script.com, cdn-4.convertexperiments.com, and conf.config-security.com. Start render stayed at 1.400s either way. The tax is visual completeness and largest paint, not the first pixel. Stores that have to keep those tools can still load them after the first screen.
Before · Scripts on
With the third parties
Speed Index 6.603s. Largest Contentful Paint 3.129s. Cumulative Layout Shift 0.002. Total Blocking Time 387ms. The filmstrip is still filling in at 3 seconds.
After · Scripts blocked
Without those three
Speed Index 2.838s. Largest Contentful Paint 2.375s. Cumulative Layout Shift 0. Total Blocking Time 352ms. The hero is in the filmstrip by 2.5 seconds.
Work with me
Is your first screen late too?
I audit Shopify pages and ship loading work that makes the first screen usable.