Most pages are slow on phones
Brent, we tested 35 of the most visited pages on a phone. Two things cause most of the slowness: product-page videos that load on their own, and pages that jump around while they load. Every fix below happens behind the scenes. Tap Fix it on anything you approve.
The site looks exactly the same
Every fix is behind the scenes. Shoppers see the same site they see today, only faster.
Videos load before they're seen
Two autoplay videos and eight clips load the moment a framed product page opens, even though they sit far below the buy box.
The page jumps
Price, Add to Cart and the variant list arrive late and push the page around. Google's limit is 0.1.
See what movesToo many apps
About 15 apps and trackers load on every page before the shopper can tap anything. Several do the same job.
See every appWhat moves on Out Of Office, second by second
We recorded every movement while the page loaded on a phone. Each bar is one jump; taller means more of the screen moved. A shopper who starts reading or reaches for a button at these moments hits the wrong spot.
- 1.0 sThe whole right column appears late. Title, price and Add to Cart are missing at first, then pop in 468 pixels tall and push the page down.0.59
- 1.1 sAdd to Cart moves. The Club Price line ("Club Price: $xx, Join Free") is added under the button after the page shows, so the button shifts.0.25
- 1.3 s and 2.3 sThe page slides up and down about 100 pixels, twice. The variant list for the 82 frame and size options is built, then rebuilt.0.27
- 6.4 sThe Club pop-up opens and nudges the page.0.10
The fix: hold the exact space for each piece from the first moment, so everything appears in its final spot instead of sliding in. The finished page looks the same as today. Unframed pages show the same pattern (0.41 to 0.60).
Every app that loads on a product page
Each app adds code the phone has to download and run before the page responds. We keep every app shoppers see. We only change when they load, and remove trackers that run twice.
| App | What it does | Speed cost | Shoppers see it? | Our plan |
|---|---|---|---|---|
| Boost | Collection filters and search | 1.5 to 4 s to start | Yes | Replace with our own Search and Filter. It tested faster on every collection. See below and fix 2. |
| Microsoft Clarity | Records visitor sessions | 0.6 s, up to 3.7 s on unframed pages | No | Two recorders do the same job. Keep one, your pick. |
| Hotjar | Records visitor sessions | 0.25 s | No | |
| Google tags | Analytics and ads tracking, loaded 3 times | 593 KB | No | Load once. Same tracking, same reports. |
| Shopify pixels | Meta, Google and Klaviyo tracking | 1 to 1.4 s | No | Keep. Start after the page shows. |
| Wiser | "You may also like" upsells | 0.6 s, plus 263 KB inside the page | Yes | Keep. Start after the page shows, in its usual spot. |
| Rebuy | Upsell and cart offers | 311 KB | Yes | Keep. Start after the page shows. |
| Upcart | Slide-out cart | 152 KB | Yes | Keep. Start after the page shows. |
| Globo | Product options | 146 KB, plus 139 KB of settings in the page | Yes | Keep. Move its settings out of the page. |
| Countdown Timer | Sale countdowns | 129 KB of timer settings on every page | On sale pages | Keep. Load only on pages that show a timer. |
| accessiBe | Accessibility widget | 247 KB | Yes | Keep. Start after the page shows. |
| Klaviyo, Okendo, Meta | Sign-ups, reviews, ads | Small | Yes | Keep as is. |
Wiser, Rebuy and Upcart overlap. Choosing one would save more, but shoppers would see a difference, so we leave them unless you ask.
Our version is faster on every collection we tested
We ran the same phone test on four collection pages: once with Boost on the live site, once with our own Search and Filter on the test copy. Ours scored higher every time, page jumping dropped to almost nothing, and two pages that sat blank for 10 to 12 seconds showed up in about 3. We recommend switching.
| Collection | Score, Boost | Score, ours | Blank screen, Boost | Blank screen, ours | Page jumps, Boost | Page jumps, ours |
|---|---|---|---|---|---|---|
| Best Selling | 26, 18 | 36, 34 | 3.0 s | 3.2 s | 0.15 to 0.28 | 0.00 |
| Coastal | 19, 20 | 37, 30 | 2.8 s | 3.2 s | 0.30 | 0.00 to 0.10 |
| The Statement Collection | 25, 22 | 41, 35 | 11.1 to 11.5 s | 3.1 s | 0.15 | 0.00 |
| Top 20 Overall | 22, 26 | 33, 34 | 9.9 to 12.1 s | 3.1 to 3.2 s | 0.15 | 0.10 |
Each page was tested twice, so both scores are shown. Scores are red under 50, so ours is still not fast; this fix is a real step up, and the other fixes do the rest. Page jumps: Google's limit is 0.1.
Every page we tested
Score out of 100 on a mid-range phone. Green 90+, amber 50 to 89, red under 50. Tap a tag to see its fix.
Out Of Office, 36 MB on a phone
Blocking the Vimeo videos alone took it from 37 MB to 14.7 MB.
Slow for months. Videos made it heavier.
What we can do about it
What you've approved
Each fix you tap shows up here with its status. We update it as the work moves.
- Google's Lighthouse test, set to a mid-range phone on a slow mobile connection. This is the same test behind Google PageSpeed.
- Boost vs our Search and Filter: four collections, two runs each, live site vs the test copy, on the same day.
- Six pages were tested three times. The rest were tested once, so read those as a guide (about 10 points either way).
- A real phone on 4G downloads less video than the test does. The direction is right; the exact megabytes are a ceiling.
- Score gains are estimates. After the product-page fixes (videos, pop-up, 3D frame, page jumps, apps) we expect product pages to reach about 30 to 50.
Tested on the live site. Nothing on the store was changed.