How an SEO Audit Grows an Ecommerce Store's Revenue

Published 2026-10-02 · 9 min read · by the isaeo team

Why the maths is different here

On most sites, connecting an SEO fix to money takes a chain of assumptions. On a store it doesn't: a product page has a price, a conversion rate and a margin. If 300 product pages are invisible to search, you can put a number on it in one line.

The other structural difference: almost nothing on a store is a one-page problem. Products are rendered by a template, so a mistake in that template is a mistake on every product. That cuts both ways — the faults are large, and so are the fixes. One template change can restore hundreds of pages.

This is also why sampling fails on stores. A ten-page crawl reaches your homepage and top categories, which are hand-tuned and fine. The fault is in the 2,000 pages nobody looked at.

1. Products excluded by canonicals

The most expensive ecommerce audit finding, and the least visible. Variants are the usual cause: a product in four sizes generates four URLs, and the platform canonicalises three of them to the fourth — or worse, all four to the category page.

Canonicalising variants to a parent product is often correct. Canonicalising products to their category is not: it tells engines the product pages needn't be indexed at all, so they stop appearing for the long-tail queries that convert best. Nothing looks broken. The pages load, the store works, and a whole tier of demand quietly goes to competitors.

What to check: pick ten products across different categories and confirm each canonical points at itself or at a legitimate parent product. If they point at a category listing, that's your quarter's biggest fix. Canonical tags explained covers the patterns.

2. Out-of-stock pages returning 404

Products go out of stock permanently, and most platforms handle it badly: the URL starts 404ing. Every link, every ranking and every bit of accumulated authority for that page evaporates.

The right behaviour depends on the situation:

• Temporarily out of stock: keep the page live and indexable, say so on the page. Removing it and restoring it later means re-earning the ranking.
• Discontinued, replacement exists: 301 to the replacement product. Not to the category — a redirect to a broadly related page is treated as a soft 404 and the value is lost anyway.
• Discontinued, nothing similar: 301 to the parent category, or 410 if it's genuinely gone. A deliberate choice either way.

An audit finds these as internal 404s in bulk, which is the only practical way to find them — nobody clicks 2,000 products. See broken links and redirect chains.

3. Faceted URLs eating the crawl

Filters generate URLs combinatorially. Colour, size, price band and sort order across a few categories can produce tens of thousands of near-identical pages, all crawlable, none worth indexing.

Two costs. Engines spend their crawl allowance on ?colour=blue&sort=price_asc instead of your new products, so new stock takes longer to appear. And the near-duplicates compete with the clean category page for the same query.

The audit signal is unmistakable: a crawl that finds far more URLs than you have products, with parameters in most of them. The fix is a deliberate policy — decide which facet combinations deserve to be indexable (often a small set with real search demand, like colour), and make the rest non-indexable or non-crawlable. Both parts matter: noindex stops indexing but not crawling.

4. Missing product schema

Without Product markup carrying offers, your listing is a title and a snippet while competitors show price, availability and star ratings. Same position, visibly less useful — and the click goes to the result that answered the question in the results page.

The AEO version of this is now the bigger gap. When someone asks an assistant for a product recommendation with a budget, structured price and availability data is what makes your product eligible to be named. Prose that mentions the price in a sentence is not the same thing. That's the argument in AEO for ecommerce, and which schema types earn citations covers the markup.

Check that price and availability are present and accurate — stale markup claiming in-stock on a sold-out product is a different and worse problem.

Every fix is a template fix

The practical upshot for anyone auditing a store: never fix a product page. Fix the template.

The workflow that follows from that:

1. Crawl everything, not a sample. Including parameter URLs, so you can see the facet explosion.
2. Group findings by template, not by page. "Missing schema on 1,847 product pages" is one ticket, not 1,847.
3. Size each group by revenue, not by page count. 40 affected best-sellers beat 400 affected clearance items.
4. Re-crawl after the deploy. Template changes have wide blast radius in both directions — verify the fix landed everywhere and broke nothing.

Step 4 is not optional on a store. A template edit that fixes canonicals and accidentally drops the H1 has just traded one site-wide problem for another, and you'll only see it in a diff.