Last updated: July 2026
Key takeaways
- An audit runs three layers: technical, app stack, and strategy
- In our speed diagnostics, apps cause most Shopify performance problems
- Work the funnel in order: navigation, homepage, collection, product, cart
- An audit isn't a redesign. Most findings are UX changes inside your existing styles
An ecommerce site audit is a structured review of your Shopify store's code, apps, and customer data that produces a list of specific changes worth making. You can run an audit in three layers: technical, app stack, and strategy.
Run an audit before making changes
Auditing first is how you avoid expensive guesses. Most stores have a long list of things that could be improved, but from the outside, the answers are rarely obvious. A store can look dated and be structurally sound. A store can look polished and be held together by nine apps injecting scripts on every page.
One clarification worth making early: an audit isn't a full redesign. Good recommendations work within the brand and styles already on your site, and most of them are UX changes rather than visual ones. You shouldn't need a rebrand to act on one.
The three layers of the audit
The three layers are technical, app stack, and strategy, and each one answers a different question. The technical layer asks whether the code can carry growth. The app stack layer asks what the store is paying for in performance and maintenance. The strategy layer asks what shoppers actually do and where they stop doing it.
| Layer | Question it answers | Main inputs |
|---|---|---|
| Technical | Can this codebase support what you want to build? | Theme architecture, speed diagnostic, accessibility audit, code review |
| App stack | What stays, what goes, what can be built natively? | Installed app list, script load order, feature overlap, cost |
| Strategy | Where is revenue leaking, and why? | Questionnaire, Shopify analytics, GA4, heatmaps, past tests |
The layers run in parallel, not in sequence. They also correct each other, which is the point of running all three. A strategy finding that says "the product page underperforms" reads differently once the technical layer shows that page carries four review widgets.
Layer 1: the technical audit
The technical audit covers codebase health: the state of the theme, where the technical debt sits, what's causing performance problems, and whether the store runs on current Shopify architecture. It has three parts. A site speed diagnostic, an accessibility audit, and a general code review.
The speed diagnostic exists to identify the root cause. That single finding changes what every other recommendation should say.
PageSpeed Insights separates Insights from Diagnostics. Start at the top: render-blocking requests here cost an estimated 890 ms, which is the kind of single finding that reorders everything below it.
Speed is worth the hour. Google and Deloitte Digital's 2020 study Milliseconds Make Millions tracked 37 brands and found that a 0.1-second mobile improvement lifted retail conversions 8.4% and AOV 9.2%. Numbers like that are why we put so much effort into Shopify speed optimization.
Architecture is the next check. Shopify's documentation on theme architecture versions sorts themes into vintage, Online Store 2.0, and the current generation that supports theme blocks. Open your theme's templates folder: a product.liquid file instead of product.json means legacy architecture.
The accessibility audit runs at the same time, because retrofitting costs more than building it in. Fun fact: WebAIM's 2026 analysis of the top one million home pages found detectable WCAG failures on 95.9% of them, up from 94.8% in 2025. This portion of the audit helps you join the small minority of winners.
We won't get into all the reasons why accessibility compliance is important in this article, but it's something we take very seriously around here. This ADA best practices checklist walks through how we approach these fixes.
Layer 2: the app stack audit
The app stack audit sorts every installed app into three buckets: keep it, remove it, or replace it with native Shopify functionality. The bias runs hard toward native. Apps are the primary source of both technical debt and performance problems on the stores we inherit.
The list itself is the starting point, not the verdict. A store locator that loads on every template is a different problem from one that loads on a single page.
That claim comes from our own diagnostic work rather than published research, so here's the specific version of it. When a Fuel Made speed diagnostic finds a root cause, that root cause is an app roughly 90% of the time. Usually it's a script loading on every template to power a feature that appears on one.
Four questions decide each app's fate:
- Does this app earn its load time, measured against the pages it actually affects?
- Could the same outcome ship natively in the theme, using metafields, sections, or blocks?
- Is another installed app already doing part of this job?
- What breaks on the storefront if it gets uninstalled tomorrow?
Apps that survive tend to be the ones doing genuinely hard work: subscriptions, reviews at scale, complex bundling. The casualties are usually banner apps, badge apps, and single-feature apps installed during a promotion and never removed.
Layer 3: the strategy audit
The strategy audit answers six questions about behavior, and it starts qualitative before it goes quantitative. Establish business context first: goals, product range, audience, and how long the buying decision takes. Numbers without that context get misread constantly.
Then the data. Shopify analytics, GA4 where it's implemented reliably, heatmaps, and any past A/B test results, read against these six questions:
- How are users arriving on the site?
- Where do they land?
- What do they do next?
- Where is the biggest drop-off?
- What are the trends between channels, device types, and new versus returning visitors?
- Is this an impulse purchase or a considered one?
That last question reframes everything above it. An impulse purchase wants fewer steps. A considered purchase wants more information, earlier, and often several sessions to decide.
Questions two and four in one view. Product pages take 55% of sessions here but convert at 0.8%, while the homepage converts at 4.6% on a fifth of the traffic. That gap is where the roadmap starts.
For heatmaps we default to Microsoft Clarity, which is free with no traffic limits and flags rage clicks and dead clicks automatically. Dead clicks are the underrated signal.
A desktop click map on an anonymized apparel homepage. Note how much attention the top-right icons pull relative to the navigation labels.
On Turbie Twist's mobile site, heatmap data showed a cluster of dead clicks in the top corners of the screen. Shoppers were reaching for a menu that wasn't there. Moving the mobile menu was one of several changes in that project, which resulted in a mobile conversion lift of 82%.
Drop-off analysis is also where expectations get calibrated. Baymard Institute's aggregation of 50 studies puts average cart abandonment at 70.22%, and its checkout usability research estimates that a large ecommerce site could gain around 35% in conversion rate from better checkout design alone.
Seven in ten leaving is normal. The real question is which of those seven were ever reachable, and what counts as a good ecommerce conversion rate for your specific brand.
Which pages to audit
Audit the conversion funnel in order, starting with navigation and ending at the cart. Every page has one job, and the audit is really just checking whether it does that job.
| Page | The question this page has to answer |
|---|---|
| Navigation, header, footer | Can shoppers find things the way they shop, rather than the way you organize the catalog? |
| Homepage | Does a first-time visitor understand the offer, and have a reason to believe it? |
| Collection pages | Can someone narrow a large catalog to a short list without effort? |
| Product pages | Does the page answer the questions that decide this purchase, before the buyer goes looking? |
| Cart and cart drawer | Is last-minute hesitation being addressed, or ignored? |
Secondary collection pages, secondary product templates, and campaign landing pages are worth adding during promotional periods, when a large share of traffic lands somewhere other than the homepage.
For each finding, write down five things: the recommendation itself, where the content will come from, a competitor reference if one is useful, whether it belongs in an A/B test, and any app integration notes.
Funnel order also makes dependencies visible. Say a product page recommendation assumes shoppers already know the difference between two towel materials. That recommendation changes once you can see the collection page never explains it.
The five principles every finding is judged against
Five principles decide whether something you see in the data counts as a problem worth fixing. Use these as a lens when writing recommendations.
- Understand intent before changing anything. What a shopper came to do determines whether a change helps or interrupts.
- Reduce friction, and test removing things rather than only adding them. Removal is an under-tested intervention in ecommerce.
- Repeat the information that matters. Shipping and returns policies shouldn't live on one page and nowhere else in the funnel.
- Match the experience to the decision lifecycle. A $19 repeat purchase and a $900 considered purchase need different amounts of information at different moments.
- Never trade usability for aesthetics. If a layout looks great and makes shoppers think harder, it's probably costing conversions.
Can you run this audit on your own store?
Most of it, yes, and the tools cost nothing. Shopify analytics is already in your admin. Microsoft Clarity is free. So is Google PageSpeed Insights.
You just need to look at all three layers before deciding what to change.
A condensed version to work through:
- Check your theme's templates folder for .json files. Liquid templates mean legacy architecture.
- Run your homepage and top product page through PageSpeed Insights, on mobile.
- Check the Web Performance Dashboard in your Shopify admin for real-user data over the last 30 days.
- List every installed app, then note which pages each one loads on and what would break without it.
- Run an automated accessibility scan, then check keyboard navigation and focus order by hand.
- Install Clarity, wait two weeks, then review click maps and dead clicks on mobile specifically.
- In Shopify analytics, find your top landing pages and the largest single drop-off in the funnel.
- Compare the last 30 days against the prior year, not just the prior month.
- Write down whether your product is an impulse or considered purchase, then check whether the site behaves that way.
- Order every finding by funnel stage, from navigation through cart, before you prioritize anything.
Step three lives here. PageSpeed Insights gives you a lab test on one load; this dashboard gives you 30 days of what real visitors experienced. When the two disagree, believe this one.
What to do with your findings
Order findings by funnel stage first, then by whether anything downstream depends on them. Effort is the last thing to sort by.
Anything blocking measurement comes first. You can't evaluate later changes without working analytics, so a tracking gap outranks a conversion idea. Then technical blockers that would make other work fragile, like legacy theme architecture. Then the largest single drop-off the data showed you.
Low-risk fixes can run alongside that sequence instead of waiting for it. A shipping threshold made visible in the cart, a redundant form field cut, a mobile menu label changed. We always recommend A/B testing any changes before implementing them.
Frequently asked questions
What is included in an ecommerce site audit?
A complete audit covers three layers. The technical layer reviews theme architecture, site speed, accessibility, and code quality. The app stack layer evaluates every installed app for removal or native replacement. The strategy layer combines business context with Shopify analytics, GA4, heatmap data, and past A/B test results to locate funnel drop-off.
How long does an ecommerce site audit take?
An audit's timeline depends on how many pages you cover. For the audits we do at Fuel Made, each page takes about a week. By "page" we mean a page template such as PDP, collections, or home.
Heatmap tools need at least two weeks of traffic before click maps mean anything, so install Clarity first and let it gather while you work through the technical and app stack layers.
How often should you audit your ecommerce site?
A full three-layer audit is worth running before any significant investment in the site. After that, the useful cadence is monthly KPI review rather than a repeated full audit. Compare trends against the prior period and the prior year, and reprioritize planned work when the numbers move.
Which tools do you need to audit a Shopify store?
Four free tools cover most of it. Shopify analytics and the Web Performance Dashboard are already in your admin. Microsoft Clarity provides heatmaps and session recordings with no traffic limits. Google PageSpeed Insights handles synthetic speed testing. An automated accessibility scanner covers detectable WCAG failures, though manual keyboard testing is still required.
How many apps is too many on a Shopify store?
There's no correct number, because load impact varies more than count does. The better test is per-app: does this app earn its load time on the pages it actually affects, and could it ship natively instead? In our speed diagnostics, an app is the root cause of the performance problem roughly 90% of the time.
What should you fix first after an audit?
Order findings by funnel stage, not by effort. Anything blocking measurement comes first, since you can't evaluate later changes without working analytics. Then technical blockers that would make other work fragile, like legacy theme architecture. Then the largest single drop-off the data shows. Low-risk fixes can run alongside that sequence.
Can you audit a store without analytics access?
Partially. The technical and app stack layers only need storefront and theme access. The strategy layer needs behavioral data, and without it you're guessing about drop-off. If analytics are missing or broken, fixing tracking becomes the first item on your list, and heatmap tools need roughly two weeks of collection before the data is worth reading.
Fuel Made works with Shopify and Shopify Plus brands on CRO, Klaviyo email and SMS, and custom Shopify design and development. If you'd rather not run the audit yourself, or you want a second read on what your data is showing, get in touch with our team.
