How we speed up your store
We measure your store on Google's own PageSpeed test, find the files that hold up the first screen, and fix them on a copy of your theme. Here's the method, with the real test data behind it.
Nothing changes on your store until you say so.
- Check your storeType your address at the top. Results on screen in about a minute.
- You decideFix it yourself with the tips, have us do it, or leave it. If real visitors already pass, we say so.
- Sign and pay onlineAbout two minutes, by card or PayNow.
- Fixes, on a copy firstTested on Google's machine, then live only with your written OK.
Why your PageSpeed result matters on Google
Google uses page speed when it ranks pages, and slow pages lose buyers before they decide.
- Google ranks on speed. Google says Core Web Vitals "are used by our ranking systems". They measure how fast real visitors see your page, how quickly it answers a tap, and whether it jumps. They're the top block of every PageSpeed Insights report.
- It can decide close calls. Relevance comes first: speed won't rank a page that doesn't answer the search. When your page and a competitor's answer it equally well, the faster, steadier page has the edge.
- Slow pages lose the sale anyway. As a page goes from 1 to 3 seconds to load, the chance a visitor leaves rises by 32% (Google). A mobile site 0.1 s faster lifted retail conversions by 8.4% (Deloitte study for Google, 37 brands).
- The score shows what to fix. The 0-100 number is one simulated phone test, so it isn't what Google ranks on by itself. But a low score points at the same apps and scripts that slow real phones down, and fixing those helps both.
Sources: Google Search Central, Think with Google, Deloitte via web.dev.
Core Web Vitals: Passed
- Main content shows LCP
- 1.9 s
- Answers a tap INP
- 123 ms
- Page jumps CLS
- 0
Performance finds the files that hold real visitors up. Open Google's report
Late first frames come in whole seconds.
63 test runs of the Inlinex home page in October. The first frame came at about 0.3 s, 1.35 s or 2.4 s, never in between. No late run scored above 83.
See the numbers
| Start | First frame | Runs | Scores |
|---|---|---|---|
| On time | 0.25-0.59 s | 45 | 68-95 |
| 1 s late | 1.34-1.39 s | 5 | 63-64 |
| 2 s late | 2.31-2.50 s | 13 | 54-83 |
63 Lighthouse runs of the Inlinex home page (phone, simulated slow 4G) from one PC in Singapore, 3-30 Oct 2026.
Then everything that downloaded counts against you.
PageSpeed counts every file that finished before the first frame as needed for it. So the work is simple to say: block the screen less, and download less early.
Only the page and its styles count.
before the late frame. On Google's machine on 12 Oct that was 124 requests, scored as "main content at 11 s".
What we change
Each fix went live on Inlinex first, after testing it on Google's own machine.
- App stylesheets load in the background
A chat app's stylesheet was the usual trigger of the late start.
home page 87 → 99-100 with the next two - Chat and session recording wait for the first touch
They start the moment someone taps or scrolls, or 4 s after the page loads.
Clarity alone was adding ~1 s - Cart drawer styles load in the background
Nobody sees the drawer until they tap the cart, so the first frame shouldn't wait for it.
- Videos load when scrolled near
One attribute on the video frames in product descriptions.
product page 57 → 86 - Header and headline in one frame
The home page's first frame becomes its whole first screen, not two half-frames.
- Links skip the redirect
Each hop costs about a second in Google's test, and real visitors wait for it too.
16 menu links fixed
Tried on Google's own test. Didn't work.
These still turn up in speed guides. We won't charge you for them.
- Render-blocking until the headlineWorked locally, then made Google's test start late 4 times out of 4.
- Loading Google Fonts after the pageThe late first frame came after the load event anyway.
- An extra preconnect hintOne on-time run out of two: no better than without it.
- A tiny animation to keep frames comingLooked good once, then failed in fair side-by-side runs.
- Hiding product pages until the photo loadsHelps a text headline, not a photo. Scores stayed at 86 and 81.
- Judging a change on one runThe same page scores 65 or 85 depending on the start. We compare 4 or more.
See what's holding up your store.
The free check reads your store in about a minute. Then decide: fix it yourself with the tips, or have us do it from S$350. See prices or read the case study.