What a speed audit of your Shopify store actually checks

A speed report usually arrives as a score and a list of warnings in technical English. It helps little if nobody knows what each line means or where to start. Here is what a speed audit really checks on a Shopify store, and how to read the result without getting lost.

What is measured, and from where

Before looking at any number you need to know where it was measured from. There are two sources and it pays not to mix them:

• Lab data: a controlled test, with a simulated phone and connection. It is repeatable and useful for finding the cause of a problem.
• Field data: what real visitors experienced, on their own phones and connections. It is what Google uses to judge the experience, and it only exists if the store gets enough visits.

Your computer and your wifi are not your customer's average phone. A store that flies on your desk can crawl on a mid-range phone with mobile data, so the review always starts from the phone.

The three signals that matter

Google summarizes the loading experience in three signals, the Core Web Vitals. Each one answers a question your customer has:

When the main content appears (LCP)

It is the moment the largest element of the first screen is painted, almost always the main photo or the headline. Google considers it good up to 2.5 seconds. The usual causes are an image that is too heavy, a font that blocks the text, or a script that runs before the page paints.

How fast it responds to a tap (INP)

It measures the time between a tap and the visible response, for example when opening the menu or choosing a variant. Good is up to 200 milliseconds. It is almost always slowed down by JavaScript from apps and tracking scripts competing for the same browser thread.

How much the page moves while loading (CLS)

It is the jump of the content when something shows up late: a banner, an image without a size, a font that changes the width of the text. Good is up to 0.1. It is the easiest signal to fix and the one people notice most: nobody wants to tap a button that moves.

What weighs underneath

Behind the three signals, an audit checks the usual suspects:

1. Installed apps. Each one can load its own code on every page, even where it does nothing. Sometimes leftovers remain from apps that were already uninstalled.
2. Images. Format, real dimensions against the ones displayed, and whether the main image loads first while the others wait.
3. The theme. Code nobody uses, animations that compute non-stop, or whole libraries for a single effect.
4. Third-party scripts. Tracking, chat, reviews and ad pixels. We check which ones are needed and when they can load.
5. Fonts. How many families and weights are requested, and what is visible while they arrive.

How to read the result

A good report does not sort by score but by the effect on the person buying. A practical way to read it:

1. Start with the pages that get the most visits: the home page, the main collections and the most viewed product page.
2. Separate what comes from your code from what comes from apps and third-party scripts. The first gets fixed; the second gets negotiated, postponed or removed.
3. Prioritize what is cheap and visible: images without a size, a main photo that loads late, a banner that pushes the content.
4. Measure again after each change and compare using the same method. If you change two things at once you will not know which one worked.

A full audit includes talking with whoever knows your operation: which apps are essential and which are not. If you want to do it with your store open, book a diagnostic.