Shopify provides the infrastructure, CDN and part of the optimization of the served files. However, the resulting performance is affected by the chosen theme, custom code, apps, analytical tags, images, videos and the amount of content on the page. Optimization therefore begins with measuring a specific problem.

Core Web VitalsRecommended limits · 75th percentile
LCP≤ 2.5 sLoading the main content
INP≤ 200 msResponse to interaction
CLS≤ 0,1Visual stability of the page
Google Web VitalsBoundaries are evaluated separately for mobile and desktop visits.

The three Core Web Vitals metrics

MetricWhat is he watching?Good value
LCPHow long does it take to display the main content of the page2.5 seconds at most
INPHow quickly the site responds to interactionsMaximum 200ms
CLSHow much the content scrolls unexpectedly when loadingNot more than 0.1

Rating refers to the 75th percentile of visits. So it is not enough for a fast developer laptop to achieve a good result. Shopify works with real visits in its reports and also separates devices and site types.

Real data versus Lighthouse

Performance reports in Shopify use data from real visitors. These capture different devices, networks and behaviors, but need sufficient traffic and can be delayed. Lighthouse creates a laboratory test under defined conditions. PageSpeed ​​Insights can also add real Chrome UX Report data to it if it has enough data for that page or domain.

Therefore, the results may not match. A lab score is not a business goal in and of itself. More important is the trend of real data, problematic site types and the impact of a particular change.

The most common sources of slowdown

  • Theme and custom code: large JavaScript packages, blocking scripts, complex sections, or code loaded even where it is not needed.
  • Apps: embedded scripts, widgets, recommendations, chats or analytics may add network requirements and work for the browser.
  • Media: too large images, missing responsive variants, automatic videos or unspecified dimensions degrade both LCP and CLS.
  • External services: tag managers, ad pixels and other libraries can mainly affect INP.
  • Page content: an excessive number of sections, products or carousels lengthens the rendering and operation of the device.

A practical order of optimization

01
Record the default state

Save real-world metrics, lab tests, and last modified date.

02
Find the problematic page type

Measure the homepage, collections and products. Each type can have a different cause.

03
Specify a specific element

Check the main image, scripts, apps, layout offsets, and long JavaScript tasks.

04
Change one cause group

The result can then be attributed to a specific adjustment instead of a random set of changes.

05
Compare the period after the change

Wait for new real-world data and track functionality and business metrics as well.

Optimisations that often make sense

  • Do not delay loading images visible without scrolling the page
  • Use responsive images with correct dimensions
  • Load non-critical functions only on interaction or later
  • Restrict JavaScript where HTML and CSS can handle the same function
  • Remove unused apps, embedded code and tags
  • Reserve space for images, banners and dynamic elements
  • Test theme changes on real page types and on mobile

Not every feature has to disappear

Some element may degrade performance, but at the same time be important for purchasing decisions. The goal is not to remove all apps, videos, or analytics. It is necessary to compare their benefit, technical price and the possibility of easier implementation.

The fastest site without information may not sell. Good optimization seeks a balance between content, features and responsiveness.

What should a performance audit include?

The audit should combine real metrics, a test of the main themes, a list of apps and scripts, and a specific order of fixes. A general export from PageSpeed ​​Insights without knowledge of the Shopify theme is usually not enough to implement safely.

Frequently asked questions

What speed score is good for Shopify?

One universal number is not enough. Mainly watch LCP, INP and CLS in real data and their development. Lab scores are useful for diagnosis, not as a sole business objective.

Do Shopify apps always slow down an online store?

Not every app has the same impact. It decides what it loads in the customer-facing storefront, on which pages and when. The specific impact needs to be measured.

Why is the online store slower on mobile than on a computer?

Mobile devices may have a weaker processor, slower network, and less room for complex JavaScript or large media. Therefore, the mobile experience needs to be measured and tested separately.

Shopify performance audit

Is your store particularly slow on mobile?

Send me the store URL and a description of recent changes. I will review the theme and loaded resources, then prioritise improvements by impact and risk.

Request a performance audit →︎

Resources and further reading