● Findings · October 2026 · Daniel

How “guaranteed 90+ PageSpeed” Shopify services actually work

Look for help with a slow Shopify store and you’ll find plenty of offers promising a PageSpeed score of 90 or more, often for under $150. That’s a bold promise. Of 1,374 Shopify stores we tested this month, only 2.1% scored 90+ on mobile. The median was 61.

So how do they get there? Sometimes with real fixes. Sometimes with tricks that make the test look better without making the store any faster for the people who shop on it. We read the homepage code of 1,372 of those stores to see which tricks are actually out there, and how common they are.

First, what PageSpeed is measuring

• PageSpeed Insights loads your page once, on a simulated mid-range phone with a slowed-down connection, and scores what it sees.
• It doesn’t scroll, tap or type. It just waits.
• The score for the same page moves by 10 points or more between runs.
• It’s a lab test. It’s great for finding problems, but it isn’t what Google ranks you on. For search, Google uses Core Web Vitals measured from real Chrome users visiting your store.

The gap between that lab test and your real shoppers is where all the tricks live.

Trick 1: Show the test a different page

A script on the page checks whether the visitor is the testing tool. That’s easy to tell: PageSpeed’s browser calls itself “Chrome-Lighthouse”, and its servers run on Linux, which very few shoppers use. If the visitor looks like the test, the script stops other scripts from loading: email popups, Google tags, ad pixels, review widgets, app code.

The test sees a light, nearly empty page and scores it high. Shoppers still get every script, exactly as before.

What we found: 3 of 1,372 stores (0.2%). Two switch off or hold back other scripts (email marketing, Google tags, ad pixels, app code) when the visitor’s machine reports Linux. One theme skips its main stylesheet when the visitor is Lighthouse.

Nothing gets faster for shoppers, and the score in front of you stops telling you what’s actually slow. It also hides the real problems from whoever you hire next.

Trick 2: Hold every script until the shopper touches the page

The page loads with almost no JavaScript. Everything else waits until the first scroll, tap, key press or mouse move. PageSpeed never does any of those, so it measures almost no script work, and blocking time drops close to zero.

What we found: 25 of 1,372 stores (1.8%), mostly from speed apps with a “delay JavaScript” setting, plus a hand-written version that gets copied from store to store.

Unlike trick 1, this is a real change. But the bill goes to your shoppers:

• The first tap on add to cart or a size option can do nothing, because the code behind it hasn’t loaded yet.
• Chat, reviews and popups show up late, sometimes after the shopper has already scrolled past.
• Ad and analytics pixels miss visitors who leave without interacting, so your ad data gets worse.

Done carefully, for a few scripts that really aren’t needed early, delaying is fine. Applied to everything, it mostly moves the wait from the test onto your customers.

Trick 3: Take things out

Remove apps, swap the slideshow for a still image, drop the video. The score goes up because the page does less. Sometimes that’s the right call, but it should be your decision, made knowing what you’re giving up, not a side effect of a score guarantee.

Trick 4: Pick the number

Run the test ten times and screenshot the best one. Report desktop instead of mobile (90+ on desktop is easy). Test the homepage only. Nothing about the store changed. It’s just a better photo.

How to check your own store

1. Look at the other half of PageSpeed. At the top of a PageSpeed report, the section “Discover what your real users are experiencing” shows data from real Chrome users. If the lab score is great but the real-user section is failing, the score isn’t telling you the truth about your store. Smaller stores may not have enough traffic for this section to show.
2. Search your page source. Open your store, view the page source (Ctrl+U on Windows, Option+Cmd+U on Mac), and search for “Lighthouse” and “Linux x86_64”. A mention inside a list of bots, used to keep them out of your analytics, is normal. A check that’s followed by code that blocks scripts or styles is the red flag.
3. Tap right away. Open your store on your phone on mobile data, and tap add to cart or open the menu the moment the page appears. If the first tap does nothing, scripts are being held back.
4. Run our free check. It flags both tricks, and tells you when the scores in the report itself are inflated because of them.

What a realistic goal looks like

Most stores with a normal set of apps score somewhere around 50 to 70 on mobile, and plenty of them sell well. A better goal than a number: the first screen shows up fast, the page responds to the first tap, nothing jumps around while it loads, and Google’s real-user Core Web Vitals pass. That’s what shoppers feel, and what Google ranks on.

When you hire someone, ask them to measure before and after the same way (same pages, several runs, mobile and desktop) and to show you exactly what they changed. If they promise a number before they’ve looked at your store, ask them how.

How we got these numbers

In October 2026 we ran Google PageSpeed Insights once on mobile for 1,374 public Shopify stores, and fetched the homepage HTML of 1,372 of them. We counted only patterns we could confirm by reading the code, and checked every trick 1 match by hand. Code that only filters bots out of analytics wasn’t counted. Tricks that live in separate script files don’t show up in homepage HTML, so the real numbers are likely higher. We don’t name any of the stores.