Practical guide · Online stores

Ecommerce website audit: from product page to checkout

By Antoine Le Louët · Estimated reading : 9 min · References

An ecommerce audit checks whether visitors can find a product, understand the offer and complete their order. This checklist helps you examine each step, keep evidence and choose useful fixes.

The method works for Shopify, WooCommerce, PrestaShop and custom stores. It combines technical checks on public pages with manual shopping journey tests. A page score cannot assess the entire store or explain its sales.

Choose the pages and scenarios to test

What to check: List the homepage, an important category, a popular product, a product with variants and an out-of-stock item. Include the cart and checkout steps. Review this sample on mobile and desktop as a signed-out visitor. Record the date, device and cookie consent state.
Evidence to keep: Keep an inventory of URLs and scenarios with an expected outcome for each: finding an available size, changing a quantity or selecting a delivery location. Distinguish tested pages from templates and products you have not checked yet.
What to fix: Start with journeys that matter to your business. If you have analytics, use landing pages and the steps where visitors leave to choose your sample. A conversion drop alone does not prove a technical fault.

Check that categories and products can be found

What to check: Follow links from the menu to a category and then a product without relying solely on internal search. For pages intended for Google, check HTTP status, robots directives, canonical and sitemap. Also examine URLs generated by filters, sorting and variants.
Evidence to keep: For one category and one product, keep the final URL, canonical tag, any noindex directive and incoming internal links. If you have Search Console access, compare with URL Inspection: an accessible page is not necessarily indexed.
What to fix: Restore missing links to important products and fix accidental exclusions. Decide which filtered pages have their own value before changing indexing rules. Do not make every sort order and filter combination indexable by default.

Make product information consistent and understandable

What to check: Read the name, description, price, availability, variants, delivery and returns information. Change a size or colour: the image, stock, price and item added must stay consistent. Check alternatives for informative images and test Product structured data with Google’s Rich Results Test.
Evidence to keep: Keep a screenshot of the selected variant and its contents in the cart, the information actually displayed and the structured data test result. Compare the markup with the visible price and availability.
What to fix: Fix conflicting information and unclear selectors before adding secondary details. Align structured data with the page. Valid markup may make a product eligible for certain rich appearances; it guarantees neither their display nor a ranking.

Measure mobile speed on sales pages

What to check: Test a category and a product separately with PageSpeed Insights or Lighthouse. Observe the main image, layout shifts and how filters or the add button respond. Compare repeated tests under the same conditions with field data when available.
Evidence to keep: Keep the URLs, test conditions and dates. Field targets are LCP ≤ 2.5 s, INP ≤ 200 ms and CLS ≤ 0.1 at the 75th percentile, separating mobile and desktop. If field data is unavailable, say so: a Lighthouse page-load test does not measure real-user INP.
What to fix: Optimise oversized images first, reserve their space and examine scripts from apps, chat and reviews. Measure after each fix. An improved technical score cannot promise an increase in revenue.

Test adding to and updating the cart

What to check: Add a product and a variant, change the quantity, then remove an item. Check stock limits, displayed charges and total recalculation. Try an invalid discount code and returning from checkout to the cart.
Evidence to keep: For each failure, record the exact steps, product and variant, device, expected outcome and actual result. Keep a useful message or network error without recording personal data.
What to fix: Prioritise incorrect variants, inconsistent totals and actions that do nothing. Make add-to-cart confirmation perceivable and explain stock or discount failures. Repeat the scenario after the fix with a fresh session.

Walk through checkout and payment, including by keyboard

What to check: In a test environment, place an order using your provider’s test payment methods. Try guest checkout if available, an invalid address, a declined payment and a successful payment. Repeat using a keyboard: visible focus, named fields, understandable errors and access to the final button.
Evidence to keep: Record where the journey stops and whether entered information is preserved. Check that success produces a single order and a matching confirmation. On the live store, stop before placing an order or charging a payment unless you have an established internal procedure.
What to fix: Fix delivery, form and payment blockers before visual refinements. Associate error messages with the relevant fields and provide a useful way forward after a decline. An automated accessibility or security score cannot validate this journey or certify the store.

Turn findings into a verifiable action plan

What to check: Rank each finding by blocked task, affected pages, observed frequency and effort to fix. Separate reproduced failures from hypotheses. If you analyse abandonment, first check that measurement events are neither missing nor counted twice.
Evidence to keep: For each action, keep the URL or scenario, evidence, observed impact, owner, deadline and acceptance criterion. Use the same conditions for before and after comparisons, then monitor errors and journey steps after release.
What to fix: Schedule order blockers first, then the most widespread discovery and usability problems. A score helps identify topics; your evidence and business context determine their order. Repeat affected scenarios after a theme, app or payment change.

Prioritisation example for a fictional store

You reproduce a checkout button that cannot be used by keyboard, an important category accidentally marked noindex and an oversized main image. One possible order, without assuming any sales uplift, is:

Unblock keyboard checkout, then repeat the entire scenario with a form error and test payment.
Remove the category’s accidental noindex, check its links and request a new inspection in Search Console.
Optimise the product image and compare several mobile measurements under the same conditions.

The order may change if findings affect different pages or users. Do not infer lost revenue from a score or a cart abandonment alone.

Where does an automated Argos audit fit?

Argos analyses one public URL and gathers technical findings on performance, SEO, security and accessibility. Use a product or category page to complement your observations. The report concerns the analysed page: it does not prove the entire catalogue was checked and performs no login, cart addition, order or payment.

Ecommerce audit questions

Can you audit an ecommerce website for free?

Start with this checklist, Search Console, PageSpeed Insights and Argos’s free analysis of a public page. The main cost is your time. A complete review also needs manual cart and payment tests, access to your data and a test environment.

Do you need to audit every product page?

Start with a sample covering each template and special cases: variants, stock shortages and promotions. A theme fix may affect many pages, but does not prove every product is correct. Expand the checks if content or behaviour differs.

Does a good technical score guarantee more sales?

No. Your offer, pricing, traffic, trust and measurement also influence conversion. An audit provides findings to fix and hypotheses to test; on its own, it cannot attribute a change in sales to a particular modification.

Complete your audit

References for further checks

Start with a public page from your store

Choose a product or category page. Run its free technical audit, then add your manual cart and checkout findings to the private report.