Store performance and Core Web Vitals
Why Shopify stores are slow and what actually fixes it: Core Web Vitals, third-party scripts, and the app weight nobody budgets for.
When a store is slow the theme takes the blame. In most audits the theme is not the main cost — the accumulated third-party scripts are, each added to solve a real problem and never reviewed since.
These articles cover where the weight actually sits and which fixes move the field data rather than only the lab score.
Articles on this topic
Why GA4 and Shopify Analytics never show the same numbers — and which to trust
Orders, revenue and sessions differ between GA4 and Shopify in almost every store. The nine usual causes, which numbers to trust for what, and how to tell a normal gap from a tracking bug.
How to see what apps a Shopify store uses — and what they cost in speed
A detector, the page source and the Network tab show most of a store's app stack. What they reveal, what they miss, and how to price each app in page speed.
Core Web Vitals for Shopify: The Technical Playbook
LCP, CLS and INP fixed one metric at a time, with the specific Liquid and JavaScript changes that move each number.
Why your Shopify store is slow: eight real causes
The theme is rarely the culprit. App scripts, leftover code, images, fonts and tracking tags — what each one actually costs you.
Third-party scripts: the performance cost nobody budgets for
Most Shopify speed problems are not the theme. They are the accumulated scripts of apps, pixels and tools nobody has reviewed since installing them.
Frequently asked questions.
Why is our Shopify store slow?
Usually third-party scripts rather than the theme: tracking pixels, review widgets, page builders, chat widgets, and the leftovers of apps that were removed from billing but not from the theme.
Which Core Web Vitals metric matters most?
LCP is where most stores fail on mobile — how long the largest content element takes to appear. CLS and INP matter but are usually cheaper to fix.
Does a lab score predict real performance?
No. A single test load tells you what to fix; field data from real users tells you whether the fix worked. Keep the two apart.