GHOST CODE REMOVAL

194 requests removed. Load time down 96%.

Consumer Electronics Retailer · Shopify Plus

37.3s 1.5s
LOAD TIME
227 33
HTTP REQUESTS
4.98MB 847KB
PAGE WEIGHT
D (64%) A (95%)
PAGESPEED SCORE
Situation

A retailer losing conversions to invisible weight

Consumer Electronics Retailer, 2020. GTmetrix fully loaded time went from 37.3 seconds to 1.5 seconds, a 96% reduction. Requests dropped from 227 to 33; 194 removed. Page size fell from 4.98 MB to 847 KB. GTmetrix PageSpeed and YSlow scores (2020 scoring) moved from D and E to A and A. One before-and-after report pair; the report copy on file has the date, region, and browser redacted.

Years of app installs had layered JavaScript on top of JavaScript. Some of those apps were gone. The scripts weren't.

What I did

Every request audited. Active apps removed first. Ghost code deleted second.

I audited every request in the waterfall and sorted them into two categories: apps that were still active but replaceable with native code, and scripts from apps that had already been uninstalled.

The active apps came out first. A sticky add-to-cart bar was loading jQuery (86KB), a stylesheet (20KB), and two tracking pixels. That was 100KB of main thread blocking for a scroll event. I replaced it with native JavaScript using the IntersectionObserver API. Then I extracted the critical CSS for the above-the-fold viewport and inlined it directly in <head>, deferring the rest so the browser could paint immediately.

The ghost code came out next. A regex search through theme.liquid and settings_data.json surfaced remnants from page builder and review apps that were no longer installed, all still firing requests to dead endpoints. I deleted them.

Total changes: subtractive. No new apps. No optimizer tools. No redesign.

Results

194 requests out. Load time at 1.5 seconds.

Metric Before After Change
Fully Loaded Time 37.3s 1.5s −96%
HTTP Requests 227 33 −85%
Page Weight 4.98MB 847KB −83%
PageSpeed Score D (64%) A (95%) +31 pts
YSlow Score E (58%) A (90%) +32 pts
JS Deferral Score F (0) A (91) +91 pts

Revenue impact (client-reported)

  • Conversion rate: 1.2% → 2.44% (+103%)
  • Bounce rate: −47%
  • Mobile revenue: +156%
The takeaway

The 194 requests didn't need to be compressed. They needed to be deleted.

Most Shopify Plus stores accumulate this kind of weight gradually, one app at a time, over years. The monthly subscription fees are visible. The performance cost usually isn't, until it shows up in mobile conversion.

The work here was entirely subtractive. No new tools were added. No optimizer plugins were installed. The store got faster because things were removed, and because the remaining JavaScript was deferred correctly.

I apply the same audit methodology to your store. Every app mapped to its performance impact, every ghost script identified and removed.

Book a Diagnostic Audit →