Shopify CRO Checklist: 27 Technical Fixes That Actually Move Revenue

A founder came to us spending $50,000 a month on Meta Ads. Their conversion rate was stuck at 1.2%. They had already fired one media buyer. They were about to fire a second.

We opened Chrome DevTools, throttled the connection to 4G, and reloaded their homepage. The problem showed up in about 10 seconds. Fourteen different third-party scripts were loading before the page became usable. Several of them were from apps deleted months ago. One was making requests to a server that no longer existed.

The Add to Cart button appeared on screen after about 2 seconds. But tapping it did nothing for another 3 seconds because the browser was still processing that pile of dead JavaScript. On iOS Safari, the checkout page shifted as it loaded and the payment button jumped below the fold right as the customer reached for it.

The ads were fine. The traffic was good. The store was broken.

That's the problem with every Shopify CRO checklist on the internet right now. Cartylabs has 80 items. cro.media has 36. Every single one of them covers the marketing layer: trust badges, sticky headers, bestseller badges, free shipping bars. All valid. None of them go to the code level. None of them tell you how to find the scripts that are silently freezing your store. None of them mention iOS Safari as a separate failure surface. None of them give you a diagnostic sequence.

This checklist is different. It's organized as four layers, in the order you should check them. Speed first. Scripts second. Checkout third. Mobile fourth. Each item tells you what to check, what tool to use, and what number you're looking for.

Before You Start: Run These Three Numbers First

Don't touch anything on your store until you know these three numbers. They tell you which layer has the biggest problem and where to focus first.

Number 1: Your mobile CVR vs your desktop CVR. Open GA4, go to Reports, then Conversions, and segment by Device Category. Write down both numbers. If your mobile CVR is more than 1 percentage point below your desktop CVR, you have a technical problem, not a marketing problem. Mobile drives over 70% of DTC traffic. If your mobile rate is half your desktop rate, the majority of your paid traffic is bleeding out before it converts.

Number 2: Your LCP on mobile. Open PageSpeed Insights, paste your homepage URL, and run the mobile test. Find the LCP number. Above 2.5 seconds is a problem. Above 3 seconds is a serious problem. Above 4 seconds means a large chunk of your paid traffic is bouncing before the page finishes loading.

Number 3: Your checkout drop-off rate. In Shopify Analytics, go to Conversions and look at the checkout funnel. Find the step with the biggest drop. That's your most expensive leak. If the biggest drop happens at the payment step, the problem is almost always a checkout CLS issue or an iOS Safari bug, not your pricing.

Once you have these three numbers, you know which layer to fix first. Now run the checklist.

Layer 1: Speed (Items 1 to 8)

Fix this layer first. If your store is slow, nothing else on this checklist matters. A trust badge on a 5-second page doesn't help anyone. Speed is the foundation everything else sits on.

1. Measure Your LCP on a Real Mobile Connection

Tool: PageSpeed Insights (mobile test) or Chrome DevTools with Fast 3G throttle.

What you're looking for: LCP under 2.5 seconds. Target: under 1.5 seconds.

LCP stands for Largest Contentful Paint. It measures how long it takes for the biggest visible element on the page to load. On most Shopify stores, that's the hero image. Run the test on mobile, not desktop. Desktop LCP is almost always fine. Mobile LCP is where the problems hide.

If your LCP is above 3 seconds, you have a hard ceiling on your conversion rate that no amount of A/B testing will break through. The full explanation of why is in the guide on what Shopify CRO actually means at the code level.

2. Check Your TTFB

Tool: Chrome DevTools, Network tab. Look at the first HTML request.

What you're looking for: TTFB (Time to First Byte) under 600 milliseconds.

TTFB is how long your server takes to start responding after a browser sends a request. If your TTFB is above 600ms, the server is slow before any asset even starts downloading. This is usually caused by Liquid template bloat making the server work too hard to build each page. Compressing your images won't fix a slow TTFB. You need to fix the server-side code.

3. Audit Your Hero Image Format and Size

Tool: Chrome DevTools, Network tab, filter by Img.

What you're looking for: Hero image under 200KB, served in WebP or AVIF format.

Your hero image is almost always your LCP element. If it's a 1.2MB JPEG, it's costing you 1 to 2 seconds of load time on mobile. Convert it to WebP, compress it to under 200KB, and add a preload tag so the browser starts downloading it immediately instead of waiting for the HTML to finish parsing.

4. Check Your INP

Tool: PageSpeed Insights or Chrome DevTools Performance tab.

What you're looking for: INP (Interaction to Next Paint) under 200 milliseconds.

INP measures how fast your page responds when someone taps a button. This is different from load time. A page can look fully loaded and still have an INP of 800ms, meaning every button tap feels delayed and broken. High INP is almost always caused by too many scripts blocking the browser's main thread. Fix the scripts in Layer 2 and INP usually drops on its own.

5. Check Your CLS Across the Full Page

Tool: PageSpeed Insights, or Chrome DevTools with Layout Shift Regions enabled.

What you're looking for: CLS (Cumulative Layout Shift) under 0.1. Target: 0.

CLS measures how much the page jumps around as it loads. Every element that loads late and pushes other elements down is adding to your CLS score. On your homepage and product pages, CLS above 0.1 means customers are trying to tap buttons that move before their finger lands. At checkout, CLS is even more expensive, covered in Layer 3.

6. Audit Your Font Loading

Tool: Chrome DevTools, Network tab, filter by Font.

What you're looking for: Critical fonts preloaded, font-display: swap applied, no FOIT (Flash of Invisible Text).

If your fonts load slowly, the text on your page is invisible until they arrive. Customers see a blank page even though the layout is there. Add preload tags for your two most important fonts and set font-display: swap so the browser shows a fallback font immediately instead of hiding the text.

7. Check Your Image Lazy Loading Setup

Tool: Chrome DevTools, Network tab, scroll slowly through the page and watch which images load.

What you're looking for: Hero image NOT lazy loaded. Everything below the fold IS lazy loaded.

This is a common mistake. Developers add loading="lazy" to every image including the hero. That tells the browser to delay loading your most important image, which directly hurts your LCP. The hero image should load immediately. Every other image should load only when the customer scrolls close to it.

8. Set Up Predictive Prefetching

Tool: Your theme's Liquid files or a prefetch script.

What you're looking for: Checkout page starts preloading when a customer hovers over Add to Cart.

When a customer hovers over your Add to Cart button, you have about half a second before they click. Use that time. Start loading the cart or checkout page in the background so by the time they click, the next step is already partially loaded. This makes your store feel instant even if it isn't. It's one of the highest-payoff fixes on this list for the amount of code it requires.

Layer 2: Scripts (Items 9 to 15)

This is where most stores have their biggest problem. Every app you've ever installed left JavaScript in your theme. Some of it is still running right now even though the app is gone. This section shows you how to find it.

9. Open the Network Waterfall and Count Your Scripts

Tool: Chrome DevTools, Network tab, throttle to Fast 3G, reload the page.

What you're looking for: No more than 5 to 6 JavaScript files loading before your hero image appears.

Open Chrome on your desktop. Press F12. Click the Network tab. Find the throttling dropdown and change it from "No throttling" to "Fast 3G." Reload your homepage.

Watch the waterfall chart that fills up. Every horizontal bar is one file your store is downloading. Count how many JavaScript files load before your hero image appears. If you see 10, 14, or more scripts loading before your main content, you almost certainly have ghost scripts. That's the single most common finding on every store we audit. The full story of what ghost scripts do to your Shopify speed explains exactly why this matters so much on mobile.

10. Look for Red 404 Errors in the Waterfall

Tool: Chrome DevTools, Network tab, same waterfall from Item 9.

What you're looking for: Zero red 404 errors.

In the waterfall, look for any requests shown in red. These are scripts trying to contact servers that no longer exist. They're ghost scripts from deleted apps making calls to dead endpoints. Every one of them adds load time and main thread blocking to every single page load, even though they do nothing useful. These are the easiest to prove to a founder because the red rows are impossible to argue with. You can see exactly how ghost scripts affect your CVR with real before and after data.

11. Count Your Total Network Requests

Tool: Chrome DevTools, Network tab. Look at the request count at the bottom of the panel.

What you're looking for: Under 60 total requests on page load. Under 40 is excellent.

A clean Shopify store should have somewhere between 30 and 60 total network requests on the homepage. If you're above 100, you have serious script bloat. One store we audited had 847KB of dead JavaScript loading silently on every page from 3 years of app installation history. Their LCP was 5.4 seconds. After removing the dead code, LCP dropped to 1.4 seconds and CVR went from 4.3% to 10.1%. Same traffic. Same ads. Same products.

12. Cross-Reference Scripts Against Your Installed App List

Tool: Shopify admin (Apps section) and your theme files (theme.liquid, snippets folder, app blocks).

What you're looking for: Any script in your theme files that belongs to an app not in your current installed app list.

Go to your Shopify admin and look at your currently installed apps. Then open your theme editor and look at your theme.liquid file, your snippets folder, and your app blocks. If you see references to apps that aren't in your installed list, those are ghost scripts. They didn't disappear when you clicked uninstall. They're still there, still loading, still blocking your store.

13. Set a Script Load Order

Tool: Your theme's Liquid files and script tag attributes.

What you're looking for: Checkout scripts and core product functionality loading first. Analytics, chat, and marketing pixels loading last.

Most Shopify stores have no script load order. Everything fires at once and fights for the same browser resources. The right approach is simple: the scripts that make your store work come first. The scripts that measure and market come last. Add defer attributes to non-critical scripts so they don't block the page from loading. How third-party scripts slow your Shopify store covers the exact mechanism behind why load order matters so much on mobile.

14. Audit Your Cart Drawer Scripts Specifically

Tool: Chrome DevTools, Network tab. Open your cart drawer and watch what fires.

What you're looking for: No more than 4 scripts firing when the cart drawer opens.

The cart drawer is where script overload hits hardest because it's a dynamic element that triggers a fresh round of script execution every time it opens. We audited one brand that had 11 different third-party scripts firing every time a customer opened their cart. Loyalty app, upsell widget, shipping protection, review trigger, three separate tracking pixels, all running at once. On 4G the cart froze completely. The customer tapped checkout and nothing happened. We cut it from 11 scripts to 4. Cart-to-checkout rate improved the same day.

15. Remove Duplicate Tracking Tools

Tool: Chrome DevTools, Network tab. Filter by JS and look for similar script names.

What you're looking for: No duplicate analytics or tracking libraries.

It's very common for Shopify stores to have GA4 installed through Shopify natively AND through a separate script tag in the theme. Same for Meta Pixel. Two copies of the same tracking library means double the load, double the main thread blocking, and corrupted analytics data. Go through your scripts and remove every duplicate. Pick one installation method per tracking tool and delete the other.

Layer 3: Checkout (Items 16 to 21)

Checkout is where your most expensive traffic is. These are the people who already decided to buy. When checkout breaks, you lose the highest- intent customers in your entire funnel. Most stores have at least one checkout bug they've never seen because they test on desktop.

16. Measure CLS at Every Checkout Step

Tool: Chrome DevTools with Layout Shift Regions enabled, or a screen recording tool on a real mobile device.

What you're looking for: CLS of exactly 0 at every checkout step. Not 0.1. Zero.

CLS during checkout is different from CLS on your homepage. On the homepage it's annoying. At checkout it costs you sales. When a dynamic element like a trust badge, a shipping protection widget, or an upsell block loads late and pushes the payment button down, a customer reaching for Apple Pay taps empty space. Sale lost. They don't try again.

A health and wellness client had a trust badge script loading late on their checkout page. CLS was 0.31. Three times above Google's passing threshold. After we locked the DOM so nothing could move after initial load, CLS dropped to 0. iOS Safari checkout completion jumped from 24% to 39%, recovering $40,000 a month in revenue that had been invisible. The full mechanism behind how checkout layout shifts kill CVR is worth reading before you run this check.

17. Lock Your Checkout DOM

Tool: Your checkout theme files or Checkout Extensibility editor.

What you're looking for: Every dynamic element has a reserved container with a fixed min-height so nothing can shift other elements when it loads.

The fix for checkout CLS is simple in concept. If you know a trust badge, a shipping estimate, or an upsell block is going to load asynchronously, you create a container for it with a defined height before it loads. The space is already reserved. When the element appears, it fills the space that was already there. Nothing moves. Skeleton loaders work the same way. Show a placeholder box while the real content loads in the background.

18. Test Checkout on a Physical iPhone in Safari

Tool: An actual iPhone running iOS Safari. Not an emulator. Not Chrome's device mode. A real device.

What you're looking for: Every checkout step works without any element jumping, zooming, or disappearing.

This is the single most important test on this entire checklist and the one that almost nobody runs. iOS Safari handles JavaScript differently from Chrome in ways that specifically affect checkout. Payment scripts, form validation, and address autocomplete all behave differently. The bugs that show up here are completely invisible in any browser emulator or desktop test. They only appear on real iPhone hardware.

Go through the full purchase flow. Add a product. Open the cart. Start checkout. Fill in your address. Get to the payment step. Watch what happens to the page as each section loads. If anything moves, if the payment button jumps, if the page zooms in when you tap a form field, you have an iOS Safari bug that's costing you sales on your biggest mobile segment. The full list of what breaks and why is in the post on why your checkout fails on iOS Safari.

19. Fix the Input Auto-Zoom Bug

Tool: Real iPhone, Safari, any form field on your checkout page.

What you're looking for: Tapping any checkout form field does NOT cause the page to zoom in.

iOS Safari automatically zooms into form fields with a font size below 16px. When this happens, the entire page layout shifts and critical checkout buttons move off screen. The fix is one line of CSS: set your input field font size to at least 16px. It's the easiest fix on this checklist and one of the most common bugs we find on physical iPhone testing.

20. Verify Native Address Autocomplete Is Working

Tool: Real iPhone or Android device, checkout address step.

What you're looking for: Address suggestions appear as soon as you start typing your street address.

Shopify has fast, built-in address autocomplete at checkout. Many custom themes accidentally break it by overriding the input fields with custom code from apps or theme customizations. When native autocomplete is broken, customers have to type their full address manually. That's extra friction at the highest-stakes step in the funnel. Test it on a real device, not a saved-address account where autocomplete never shows.

21. Review Your Checkout Extensibility Script Load

Tool: Shopify admin, Checkout editor, installed extensions list.

What you're looking for: No more than 3 to 4 active checkout extensions. Every extension is from Shopify native features where possible, not third-party apps.

Shopify's Checkout Extensibility tools are powerful for Shopify Plus stores. But you can recreate all the same ghost script problems by stacking too many third-party checkout extensions. Each one adds JavaScript to the checkout load sequence. Audit what you have installed. Remove anything you added but don't actively use. Prefer native Shopify features over third-party apps at every point in the checkout flow. To understand the full scope of what checkout friction actually costs at the revenue level, the numbers are larger than most founders expect.

Layer 4: Mobile (Items 22 to 27)

Mobile is where everything compounds. A slow store is worse on mobile. Script overload is worse on mobile. Checkout bugs are worse on mobile. And mobile is where 70% of your paid traffic lands. These six checks are the ones that recover the most money for stores that have fixed the first three layers.

22. Segment Your CVR by Device in GA4

Tool: GA4, Reports, Conversions, Device Category dimension.

What you're looking for: Mobile CVR within 1 percentage point of desktop CVR.

This is the most important number in your analytics right now. Pull your mobile CVR and your desktop CVR separately. Most stores show desktop around 2.5% to 3% and mobile around 1% to 1.5%. That gap is not normal. It's not because mobile users are distracted. It's because mobile CPUs are weaker than desktop processors and every extra script you load takes longer to process on a phone. The gap is a technical problem with a technical fix. The full explanation is in the post on why your mobile CVR is half your desktop CVR.

23. Test on a Real Mid-Range Android on 4G

Tool: A physical mid-range Android phone on a real cellular connection. Not WiFi. Not a flagship.

What you're looking for: Every interaction feels immediate. No frozen buttons. No loading delays between taps.

Your MacBook Pro is not your customer's device. Test on the device your actual customers use. A Samsung Galaxy A-series phone or similar mid-range Android on 4G is the closest thing to your median mobile visitor. If the store feels slow or buttons feel unresponsive on this device, that's exactly what the majority of your mobile traffic is experiencing every day.

Go through the full purchase flow. Time how long it takes from tapping Add to Cart to seeing the cart drawer open. Time how long checkout takes to load. If any step takes more than 2 seconds to respond, you have a mobile performance problem that's costing you sales.

24. Test Across 15 Different Screen Sizes

Tool: Chrome DevTools responsive mode, or BrowserStack for real device testing across multiple screen sizes.

What you're looking for: No layout breaks, hidden buttons, or overflow issues on any common screen size.

A store can display perfectly on 6 device sizes and break completely on the 7th. A custom section that looks great on a large iPhone Pro Max can break on a narrower phone because the developer set overflow properties without considering how they behave on smaller viewports. Test across at least 15 screen sizes. The bugs you find at widths 375px and 360px are the ones affecting the largest share of your Android traffic.

25. Audit Your Tap Targets

Tool: Chrome DevTools, Elements panel. Or test manually on a real device.

What you're looking for: Every tappable element is at least 44 by 44 pixels. Nothing important is closer than 8 pixels to another tappable element.

Tap targets that are too small cause customers to miss buttons and tap the wrong element. This is especially bad on product variant selectors. If your size or color options are small text links or tiny radio buttons instead of large swatches, you're adding friction at the exact moment a customer is trying to commit to a purchase. Replace dropdown selectors with thumb-friendly swatches. The tap surface should be large enough that a customer can select it with their thumb without looking closely.

26. Check for a Sticky Add to Cart on Mobile

Tool: Real mobile device, product page, scroll past the Add to Cart button.

What you're looking for: A sticky bar with the product name, selected variant, price, and Add to Cart button remains visible at all times while scrolling.

On a long product page with multiple images, a description, reviews, and FAQs, a customer can easily scroll past the Add to Cart button and lose it. On desktop they can see the button out of the corner of their eye. On mobile it disappears off the top of the screen. A sticky ATC bar means the primary action is always one tap away regardless of where they've scrolled. This is one of the fastest conversion fixes to implement on mobile and one of the most consistently impactful.

27. Calculate Your Monthly Revenue Leak

Tool: A calculator and your GA4 data.

What you're looking for: The dollar gap between what your store is converting at and what your category benchmark says it should be.

This is the last item on the checklist because it's the one that tells you what everything above is worth. The math is simple.

Take your monthly sessions. Multiply by your current CVR. That's your current orders. Now multiply those same sessions by your category benchmark CVR. Multiply the result by your average order value. The difference is your monthly revenue leak.

Example: 50,000 monthly sessions, $80 average order value. At 1.5% CVR you're generating $60,000 a month. At 3% CVR (the apparel benchmark) you should be generating $120,000 a month. The gap is $60,000 a month. That's what the broken code is costing you every single month. Not a one-time cost. A recurring monthly loss until the technical problems are fixed.

To find your category benchmark and see how stores in your niche compare before and after technical fixes, the Shopify conversion rate benchmarks for 2026 has the numbers by niche with real client data attached.

How to Use This Checklist Without Breaking Your Store

Three rules before you start touching anything.

Rule 1: Never edit your live theme. Duplicate your current production theme and do all your investigation on the copy. Removing a script that looks dead can break something you didn't know depended on it. Always test on a staging theme before you touch the live store.

Rule 2: Fix the speed and script layers before the checkout and mobile layers. Ghost scripts cause slow LCP. Slow LCP causes bad INP. Bad INP causes frozen buttons. Frozen buttons hurt mobile CVR. The layers build on each other. Fixing checkout CLS on a store that still has 14 scripts loading before the page is ready is like painting a wall before the foundation is dry.

Rule 3: Measure before and after every fix. Run PageSpeed Insights on mobile before you make a change. Run it again after. Check your checkout drop-off rate in Shopify Analytics before and after fixing CLS. Pull your mobile CVR from GA4 before and after real device testing fixes. You can't manage what you don't measure, and every fix on this checklist should produce a number you can point to.

What to Do After You Run the Checklist

If you ran these 27 checks and found zero problems, your store is technically clean. Now is the right time to run A/B tests. Test headlines. Test CTA copy. Test layout variations. The data will be trustworthy because you're testing on a fast, functional store.

If you found problems, fix them before you run any tests. Running A/B tests on a store with ghost scripts and a 4-second LCP produces data that looks conclusive and is actually meaningless. You're testing two versions of a broken experience. Whichever version "wins" did so despite the technical problems, not because of the copy change. We took over from a brand that had been paying a traditional CRO agency for 6 months. That agency had been testing button colors and headline variations the entire time while LCP sat at 5.1 seconds from Liquid template bloat. When we stopped the tests, fixed the code, and dropped LCP to 1.4 seconds, more revenue came in during the first month than from 6 months of creative testing combined.

Fix the technical foundation first. Then test. That's the only sequence that produces results you can trust. The full principle behind this is covered in the complete guide to why your Shopify conversion rate is low and how to find the actual cause.

If the problems you found are too deep to fix internally, the Conversion Engineering service runs the full diagnostic, fixes the technical debt, and delivers a revenue impact number before any commitment is made. The sprint runs 21 to 30 days. Most clients see measurable CVR improvement within the first week after deployment. If speed is the primary issue, the Shopify Speed Optimization service handles the LCP, ghost script, and Liquid architecture work specifically.

Frequently Asked Questions: Shopify CRO Checklist

What is a Shopify CRO audit checklist?

A Shopify CRO audit checklist is a structured list of checks you run on your store to find the specific things stopping visitors from converting into buyers. Most checklists cover the marketing layer: button colors, trust badges, CTA placement. A technical CRO checklist covers the code layer: ghost scripts, LCP measurement, checkout CLS, iOS Safari bugs, and mobile CVR gaps. The technical layer almost always has a bigger impact on conversion rate than the marketing layer, especially for stores spending serious money on paid ads.

How do I find ghost scripts on my Shopify store?

Open Chrome DevTools (press F12), click the Network tab, set throttling to Fast 3G, and reload your homepage. Look for any red rows in the waterfall. Those are scripts making requests to servers that no longer exist. Then look at your theme files (theme.liquid, snippets folder, app blocks) and compare them to your currently installed apps. Any script referencing an app that's not in your installed list is a ghost script. Never remove scripts directly from your live theme. Duplicate the theme first, test on the copy, and only push changes live after confirming nothing broke.

How many scripts should a Shopify store have?

A clean Shopify store should have between 30 and 60 total network requests on the homepage, with no more than 5 to 6 JavaScript files loading before the hero image appears. If you're above 100 total requests, you have script bloat that's affecting your load time and your INP. If you have more than 6 scripts loading before your main content, you almost certainly have ghost scripts from deleted apps.

Does checkout CLS actually affect conversion rate?

Yes, directly and measurably. Checkout CLS causes customers to tap the wrong element or miss the payment button entirely when it moves as the page loads. A health and wellness client had a CLS of 0.31 on their checkout page. After fixing it to exactly 0, iOS Safari checkout completion jumped from 24% to 39%, recovering $40,000 a month. CLS at checkout is more expensive than CLS anywhere else on your store because it affects the customers who have already decided to buy.

Why is my Shopify mobile CVR so much lower than desktop?

The gap between mobile and desktop CVR is almost always a technical problem. Mobile CPUs are weaker than desktop processors, which means every JavaScript file your store loads takes longer to process on a phone. Ghost scripts and app bloat that are barely noticeable on desktop become serious performance problems on a mid-range Android on 4G. If your mobile CVR is more than 1 percentage point below your desktop CVR, work through Layer 2 (Scripts) and Layer 4 (Mobile) of this checklist before changing anything else.

Do I need a developer to run this checklist?

The diagnostic checks in Layers 1 and 2 (measuring LCP, counting scripts, finding 404 errors in the waterfall) can be done by anyone who can open Chrome DevTools. The fixes for ghost scripts, Liquid template bloat, and checkout CLS require a developer. Layer 4 mobile testing only requires a real phone and about an hour of your time. The most important thing is running the diagnosis first so you know exactly what you're fixing before any developer touches your theme.

Should I run A/B tests before fixing technical issues?

No. Running A/B tests on a store with ghost scripts, slow LCP, or checkout CLS produces meaningless data. You're testing two broken versions of the same experience. Whichever variant "wins" did so despite the technical problems. Fix the technical foundation first using this checklist, then run tests on a fast, clean store. The test results will be reliable and the baseline CVR you're improving from will be significantly higher.

Article by

Muhammad Usama

Founder & Head Conversion Engineer

Founder & Head Conversion Engineer with 8+ years of technical engineering experience. I bridge the gap between full-stack development and e-commerce growth, specializing in tearing down Shopify architectures, eliminating code-level friction, and building high-performance infrastructure for 7- and 8-figure brands.