Ecommerce SEO audit: the checks that come first

A useful store audit is a prioritized diagnosis, not an endless spreadsheet of tiny issues. Its job is to identify valuable pages that customers cannot find or use well.

Pick a representative sample

Select one profitable category, one large category, five varied products and one buying guide. For each URL, record its canonical address, Search Console indexing status, organic clicks and sitemap inclusion. Google recommends monitoring indexing for ecommerce sites; the same check is useful after any significant store change.

Rank problems by consequences

A best-selling product returning 404, a category blocked from crawling or a mistaken canonical matters before minor title edits. Next inspect navigation links, filter-generated duplicates and product content. A heavy image can slow the page; a missing view of the product can also prevent a confident purchase. Write down the expected benefit of each proposed fix.

Turn findings into work

Each issue needs a URL, evidence, a precise change and an owner. For example: “Hiking shoe category: 23 products can be found only via site search; link to the range from its parent category.” Try a change on a small set before applying it across a catalog.

When a URL changes, align internal links, the sitemap and canonical tags with the destination. Google explicitly recommends that consistency. Repeat the sample after deployment and compare like-for-like periods in Search Console.

Avoid a false diagnosis

A page missing from search results is not necessarily blocked. It may be new, weakly linked or too similar to another page. Inspect the URL, rendered HTML and internal links before choosing a cause. In the audit sheet, separate “observed fact,” “likely cause” and “proposed test.” That distinction prevents a store-wide redesign based on a hunch. If the same defect appears across your sample, look for the theme or import rule that generates it.

Rank findings by likely impact

An audit only helps when each finding leads to a decision. Use four columns: affected URL, evidence, effect on buying or crawling, and next action. An important product that cannot be reached through categories usually deserves attention before a minor meta description change. An issue on an obsolete page with no products or visits may wait. Assign the fix to the right owner: content, catalog, theme or development.

Illustrative example: ten products in one range display “in stock” while the selected variant is unavailable. This can erode trust and make product data inconsistent. Reproduce the issue on two products and two variants before declaring it a site-wide problem. Fix the inventory source, then check the visible page, structured data and merchant feed if one is used. A list of errors without reproduction and a later check is not an actionable audit.

A 90-minute first pass

Spend 20 minutes on the home page, two categories and their links. Use 25 minutes to follow five products through to the cart on a phone. Use another 25 minutes to check indexability, canonical URLs, titles and Search Console errors for those same pages. Reserve 20 minutes to name three fixes, their owners and a review date. This does not replace a full audit; it prevents minor issues from consuming the first week.

Choose a measure that matches each correction. A missing category link is checked by its presence and the path to the product; a price error is checked on the page and in the markup. Average ranking alone cannot prove that a technical fix worked.

A sample that exposes recurring faults

Do not audit ten random URLs from the same category. Select the homepage, two categories, three products, a guide and an old redirect. For each, record HTTP status, canonical, indexability, internal links and the answer supplied to its target search. A varied sample shows whether a problem comes from a page template, catalog process or single item. Expand the check to pages sharing the faulty template.

Compare sitemap URLs with pages reachable from menus and categories. If a product is listed in XML but no link leads to it, repair the path. If an old URL redirects to the homepage even though a matching category exists, choose a more relevant destination. If a product is indexable but contains only manufacturer text, add its own facts and evidence before relying on an SEO title.

Rank actions by observable effect

Make three columns: technical blocker, missing buying information and editorial opportunity. A server error or accidental noindex belongs in the first; an unclear delivered total in the second; a complementary guide in the third. Give every action an exact URL, owner, proof of the problem and expected result. This prevents a list of thirty recommendations without order.

Repeat tests on the same URLs after changes, then follow impressions and orders for the affected families. A newly accessible page should not be called “performing” on the same day; it has passed a technical check. The audit is done when every priority finding has a decision and a follow-up that can test its real effect.

A repeatable audit record

Create one row per sampled URL with columns for type, HTTP response, canonical, indexing directive, parent link, core buying facts and Search Console state. Add the exact date and a screenshot or HTML note for any defect. When a finding affects a shared template, list how many pages use it before estimating impact. This prevents a single awkward product page from being presented as a sitewide fault.

Assign each action an owner and a retest date. After the fix, repeat the same checks on the original URL and one sibling page. If the technical fault has gone, mark that result separately from later impressions or sales. A clean audit keeps cause, correction and outcome distinct. It also helps explain why an SEO report may show improved access before organic traffic changes.