Guide Ecommerce SEO Audit

TL;DR — An ecommerce SEO audit is a health vérifier pour votre online store — finding the choses que arrêter Google from showing votre product and category pages, and fixing the ones que matter la plupart. The trick: don’t commencer by tweaking titles and ajout keywords. Commencer dans la recherche Google Console and trouver out qui pages Google peut même find and index, parce que on a store the biggest problems hit thousands of pages at une fois.

Ce que an ecommerce SEO audit is

An audit is simplement a structured regarder at votre store to réponse un question: what’s holding it back in search? It’s the même idea as a checkup at the doctor — you’re looking pour problems avant ils cost vous, and vous fix the serious ones premier.

The raison ecommerce stores besoin leur propre kind of audit is scale. A blog has a few hundred pages. A store peut have tens of thousands — un page per product, plus a category page pour every façon to groupe les, plus a near-identical copy of chaque page pour every color, size, and sort order. That’s a lot of pages, and a lot of les regarder almost the même to Google.

Où to commencer (and où not to)

La plupart personnes commencer an audit by rewriting title tags and stuffing in keywords. That’s backwards. Si Google can’t explorer or index une page, aucun amount of keyword fonctionner va aider — lune page isn’t in the race at tout.

So I commencer in Recherche Google Console, qui is Google’s free dashboard pour site owners. The la plupart utile screen is the Page Indexation report: it indique vous qui of votre pages Google has en réalité indexé, and donne vous a raison pour every page it hasn’t. On a store, que report usually surfaces the big money problems correct away — thousands of duplicate pages, pages Google trouvé but jamais bothered to explorer, or (worst cas) pages accidentally marked “do not index.”

The two problems almost every store has

1. Contenu dupliqué. Ce is the big un. Votre filters (color, size, price) créer a brand-new web adresse every temps someone clicks un, and la plupart of ceux addresses montrer nearly the même products. Variants (the même shirt in five colors) souvent chaque obtenir leur propre page aussi. Google ends up wading via thousands of near-copies.
2. Thin category pages. A category page that’s simplement a grid of products with aucun description is hard pour Google to rank — there’s pas suffisant on it to tell Google ce que it’s à propos de.

Ce que you’ll besoin

• Recherche Google Console — free, and the unique la plupart important outil ici.
• A site robot d’exploration — something que visits tout votre pages comme Google fait and listes the problems. I utiliser Ahrefs Site Audit (ce is my site, so that’s the honest réponse), but Ahrefs Webmaster Outils donne vous a free explorer of sites vous vérifier, and Screaming Frog is a popular desktop option.
• PageSpeed Insights and the Résultats enrichis Tester — les deux free from Google — pour checking speed and votre product markup.

Vouloir the complet, prioritized checklist — every area to audit, in order, with the ecommerce gotchas and how to rank the fixes by impact? Switch to the Avancé tab.

TL;DR — I audit a store in roughly ce order: crawlability → indexation → contenu dupliqué → on-page at scale → technical (CWV, HTTPS, canonicals) → données structurées → lien internes → out-of-stock handling → off-page. I commencer in the GSC Page Indexation report and a fresh explorer, not on-page, parce que on a store the highest-leverage problems hit thousands of URLs at une fois: faceted-nav duplication, accidental sitewide noindex, “Discovered – currently not indexed” at scale, sitemaps complet of redirections. The deliverable is a prioritized impact/effort liste, pas a 500-point report. La plupart stores jamais besoin to worry à propos de budget d’exploration — but the ones que do really do.

Audit philosophy: client-first, pas issue-first

A robot d’exploration va hand vous 170+ problème types. Si vous dump tout of les into a report, you’ve produced a document, pas an audit. The job is to trouver the handful que déplacer rankings and revenue and ignore the rest.

The starting point isn’t the outil — it’s the pain. As I’ve put it avant: “Si clients are coming to vous asking pour an audit, ils déjà have a pain point. Talk to les. Solve que un chose and they’ll be happy with the audit.” On a store the pain is usually concrete: trafic dropped, a category stopped ranking, a migration went sideways, nouveau products aren’t getting indexé. Anchor the audit to que, alors expand.

And know quand to arrêter. From our study of over a million domains: “Parfois, the meilleur course of action is to ne faites pashing parce que the costs outweigh the benefits.” Pas every flagged problème is worth a developer ticket.

Pourquoi I commencer with Search Console and explorer données, pas on-page

The instinct to ouvrir the audit with title tags and headings is the la plupart courant mistake I voir. On an ecommerce site the math is contre it: tweaking a title helps un page; fixing an indexation pattern helps thousands. The biggest wins on a store are almost toujours exploration and indexation problems at scale, and the stakes are highest exactly ici — “un mistake peut garder millions of pages out of the index or supprimer an entier site from résultats de recherche.”

So the audit runs in priority order, top to bottom. Here’s the whole chose.

Step 1 — Crawlability

Peut bots reach ce que matters, and are ils wasting leur temps on ce que doesn’t?

The ecommerce-specific traps:

• Faceted navigation URL explosion. Filters (color, size, price range) combine into thousands of unique URLs. Google is candid à propos de the cost: “the robots d’exploration va typically accès a very grand number of faceted navigation URLs avant the robots d’exploration’ processes determine l’URLs are in fact useless.”
• Parameter variants — session IDs, tracking codes (?utm_source=), sort orders — multiply the même content à travers URLs.
• Explorer traps — internal résultats de recherche, infinite calendars, unbounded filter stacking.
• robots.txt over- or under-blocking — accidentally blocking CSS/JS nécessaire to render, or a valid product chemin; or failing to block a junk parameter space.
• JS-gated navigation. Google veut réel liens: “utiliser <a href> tags quand creating liens to autre content. Don’t utiliser JavaScript events on autre HTML DOM elements pour navigation.”

Audit steps:

• Récupérer and lire robots.txt; confirmer nothing important is disallowed and que low-value parameter spaces are.
• In the robot d’exploration, pull “Blocked by robots.txt” — quelconque pages importantes caught?
• GSC Statistiques d’exploration: regarder pour spikes in crawled URLs contre a flat indexé count — que gap is explorer waste.
• Export tout crawled URLs and cluster by chemin/parameter pattern to size the faceted/parameter problem.

On budget d’exploration — de-escalate premier. La plupart stores don’t have a crawl-budget problem, and I’ll dire so in the audit plutôt que invent un: “La plupart sites don’t besoin to worry à propos de budget d’exploration, but là are few cas où vous may vouloir to prendre a regarder.” It becomes real around Google’s thresholds — “Grand sites (1 million+ unique pages)…” modification weekly, or 10 000+ pages modification daily — or quand “Discovered – currently not indexed” is grand. Quand it is réel, vous fix it by removing waste, pas by asking Google to explorer plus. (Deep dive: budget d’exploration.)

Step 2 — Indexation

Crawlable isn’t indexé. The GSC Page Indexation report is the center of gravity pour the whole audit — “voir qui pages Google peut trouver and index on votre site, and apprendre à propos de quelconque indexation problems encountered.”

The statuses que matter la plupart on a store:

• Duplicate sans user-selected canonical / Google chose a différent canonical — the faceted/variant duplication problem, made visible.
• Crawled – currently non indexée — usually thin product or near-duplicate filter pages.
• Découvert – currently non indexée — Google knows l’URL but hasn’t crawled it; on grand catalogs ce is a crawl-priority signal.
• URL marked ‘noindex’ — vérifier it’s intentional. The accidental sitewide version is the “one mistake” scenario.
• Soft 404 — a 200 réponse on une page that’s effectively vide (out-of-stock, zero-result search). Google warns ces “va continuer to be crawled, and waste votre budget.”

Audit steps:

• Export “Not indexed” URLs by raison; quantify chaque bucket.
• Comparer GSC indexé count contre votre connu catalog size — a grand gap signifie Google isn’t finding pages.
• Run Inspection d’URL on a sample from chaque raison bucket.
• Audit the XML sitemap: it devrait liste seulement canonical, indexable, 200-status URLs. Redirigé, noindexed, or canonicalized-away URLs in a sitemap send conflicting signals — pull les.

Step 3 — Contenu dupliqué (the #1 ecommerce problème)

Ce is the unique la plupart courant chose an ecommerce audit uncovers, and it’s worth its propre step. Google’s Gary Illyes has estimated que roughly 60% of the internet is contenu dupliqué — and stores are overrepresented.

Où it comes from:

• Faceted navigation — ?color=blue, ?size=S&color=blue, ?size=S&color=blue&sort=price tout serve near-identical content.
• Parameter-order permutations — ?color=blue&size=S and ?size=S&color=blue are the même page twice.
• Product variants as separate URLs — ?variant=… sans a self-referencing canonical or a canonical to the parent.
• A product living sous multiple category paths, les deux indexable.
• HTTP/HTTPS, www/non-www, trailing-slash inconsistency que doesn’t 301 to un canonical.

The Shopify gotcha worth appel out by nom. Shopify does canonicalize /collections/{collection}/products/{product} to the clean /products/{product} URL — so personnes assume it’s handled. It isn’t entièrement: the lien internes from collection pages encore point at the non-canonical /collections/... version, so the canonical product URL receives aucun internal popularité des liens via normal navigation, and it peut montrer up as an orphan. Fixes: ajouter an “All products” page que liens the canonical /products/ URLs, or edit collection templates to lien the URL canonique directement.

Audit steps:

• GSC: pull les deux “Duplicate” buckets from Page Indexation.
• Robot d’exploration Content Quality / duplicate-cluster report: trouver clusters with aucun canonical specified.
• Confirmer faceted parameter URLs are handled — Google’s recommended prevention is robots.txt blocking of filter parameters, or fragment-based (#) filtering, qui “will have no impact on crawling.” Don’t reach pour noindex ici (voir Myths).
• Spot-check que Shopify /collections/*/products/* URLs canonicalize to /products/* — and que something liens the canonical version.

(Background: contenu dupliqué and canonicalization.)

Step 4 — On-page at scale

On a store, on-page is a templating problem, pas a copywriting un — you’re auditing patterns, pas pages.

The recurring problèmes:

• Thin category pages (the #2 ecommerce problème). John Mueller is direct à propos de it: “Quand the ecommerce category pages don’t have quelconque autre content at tout, autre que liens to the products, alors it’s really hard pour us to rank ceux pages.” But don’t overcorrect into a wall of text — “maybe 90%, 95% of que text is unnecessary,” and “our algorithms parfois obtenir confused quand ils have a liste of products on top and essentially a giant article on the bottom.” The target is a short, utile block (sourcing, materials, sizing, popularity) near the top — pas filler que buries the grid.
• Templated title collisions — Buy {Product} | Store patterns que go near-identical à travers products, or two products sharing a nom.
• Manquant/vide H1s and titles at the template level.
• Meta descriptions — worth doing on votre top category and product pages, moins worth it on the long tail (Google rewrites les la plupart of the temps anyway).

Audit steps:

• Robot d’exploration On-Page report: filter pour manquant title, manquant H1, manquant meta description, duplicate titles, duplicate H1s.
• Flag category pages with little to aucun unique corps copy (thin-content candidates) and cross-reference contre “Crawled – currently not indexed.”
• Export titles and vérifier pour template collisions.

A remarque on ce que not to spend audit temps on: multiple H1s. They’re valid HTML5 and near-irrelevant to ranking — 51,3% of sites have les somewhere, per my propre million-domain study. Skip it.

Step 5 — Technical (CWV, HTTPS, canonicals, mobile)

Core Web Vitals. Google’s passing thresholds: LCP < 2,5s, INP < 200ms, CLS < 0,1. Ecommerce échec modes are predictable — unoptimized hero images and third-party scripts (chat, A/B, pixels) hurt LCP; add-to-cart and checkout JS hurt INP; images sans explicit dimensions and injected promo bars hurt CLS. Audit by URL groupe in the GSC CWV report so vous pouvez voir si it’s product, category, or checkout templates failing, alors confirmer on representative templates in PageSpeed Insights. (Voir Core Web Vitals.)

HTTPS. Everything — product, cart, checkout — on HTTPS, aucun mixed content (HTTP images/resources on HTTPS pages), and canonicals/redirections pointing at the HTTPS version.

Canonicals. The courant ecommerce errors: canonicals pointing at 4XX pages; non-URL canoniques in le sitemap; paginated pages canonicalized to page un (don’t); and — the quiet un — lien internes pointing at non-canonical versions so the canonical jamais accumulates popularité des liens. Multiple rel=canonical tags on un page obtenir ignored.

Mobile-first. Google indexes the mobile version. Assurez-vous the mobile product page isn’t a stripped-down un manquant the description, the données structurées, or full-size images.

Raw versus rendered product evidence. On representative PDPs, enregistrer the initial HTML and a rendered capture après a fresh navigation. Ne faites pas treat resizing an already-loaded desktop page as a mobile-rendering tester. The raw réponse devrait expose the core product identity, selected/par défaut SKU and attributes, price, currency, availability, crawlable variant liens où requis, and matching Product/Offer données. Alors comparer the rendered DOM and visible selection. Google may render JavaScript, but autre robots d’exploration and agents vary; a successful render aussi ne fait pas cure contradictory product facts.

Pour chaque sampled variant, continuer the comparison via the feed, cart, and checkout. L’URL, visible PDP, rendered JSON-LD, feed item, and transaction devrait agree on product identity, SKU, groupe ID, selected attributes, price, currency, and availability. Record intentional emplacement- or customer-specific revalidation separately from unexplained mismatches. Voir Product Page SEO pour the raw/rendered boundary and Product Variant SEO pour the complet selected-offer contract.

Step 6 — Données structurées

Pour a store, the priority types are Product / ProductGroup, BreadcrumbList, Organization (with retourner policy), and Examiner / AggregateRating. Ajout plus valid properties widens eligibility — Google: “ajout the plus properties vous pouvez ajouter, the plus enhancements votre page peut be eligible pour.” (The un myth to garder in votre head: schema rend vous eligible pour résultats enrichis; it doesn’t faire vous rank.)

Audit steps:

• Run the Résultats enrichis Tester on a representative product and category page.
• GSC Enhancements: vérifier Product Snippets and Merchant Listings pour errors/warnings.
• Vérifier requis vs. recommended separately per fonctionnalité — don’t collapse les. Pour merchant listings, Google’s requis définir is name, image, and offers (an Offer, with price, priceCurrency, and availability). Pour product snippets, name is requis and vous besoin au moins un of review, aggregateRating, or offers to be eligible — Google listes aggregateRating, offers, and review as recommended, pas requis, on que fonctionnalité.
• Courant faults: manquant image, manquant offers, manquant availability, schema prices que don’t match the visible price (a policy violation), breadcrumb schema que doesn’t match the visible trail.

Step 7 — Maillage interne

The problèmes: orphan product pages (classic Shopify symptom, ci-dessus); important category/product pages buried 5+ clicks deep; lien internes pointing at redirections or 404s; meilleur sellers pas lié from high-authority hubs. Google: “ajouter liens from menus to category pages, from category pages to sub-category pages, and finalement from sub-category pages to tout product pages.”

Audit steps:

• Robot d’exploration Liens report: pull orphan pages and explorer depth; confirmer clé pages sit dans ~3 clicks of the homepage.
• Pull liens to redirections and broken liens.
• Vérifier que votre highest-revenue products are lié from the principal nav and editorial/hub pages.

Step 8 — Out-of-stock and discontinued products

A genuinely ecommerce-only section. My honest réponse ici is the SEO cliché pour a raison: “though it’s a joke in the SEO community, ‘it dépend’ is really the réponse quand dealing with out-of-stock products.” And “ultimately, there’s aucun perfect solution.” The decision turns on si it’s temporary vs. permanent and si lune page has trafic or liens:

Ne faites pas audit availability from the visible étiquette alone. Sample an ordinary in-stock SKU, a temporary stockout, a recently restocked item, un unavailable variant à l’intérieur an disponible product groupe, and a postcode-restricted offer. Pour chaque, reconcile the authoritative backend state, selected variant, visible page, Product/Offer markup, merchant feed row, cart line, and checkout outcome. Record the market, postcode, channel, collection temps, and feed-processing temps so a personalized fulfillment result n’est pas mistaken pour the catalog’s general state. The complet mapping is in the Product Schema availability contract.

Watch pour soft 404s: out-of-stock pages, zero-result searches, or vide cart/account pages returning 200. Pull the Soft 404 bucket in GSC and resolve by pattern. (Voir out-of-stock products pour the complet framework.)

Step 9 — Off-page

Lighter pour la plupart stores, but worth a réussir:

• GSC Performances: branded vs. non-branded click share — heavy branded reliance signifie weak non-branded discovery.
• Backlink profile by page type — are category/product pages earning liens, or seulement the homepage and blog?
• Trouver 404 pages que have backlinks and 301 les to the closest live page (reclaim the equity).
• Competitor lien gap on category pages.

How to prioritize the findings

Everything ci-dessus produces a liste. The liste n’est pas the deliverable — the sorted liste is. I score chaque finding on an impact/effort matrix: “anything high-impact and low-effort is a rapide win, so ceux tasks devrait be tackled premier.” Pour a store:

• Élevé impact / low effort — do premier: a stray sitewide noindex; a sitemap complet of redirection/noindex URLs; manquant canonicals on variant URLs; lien internes pointing at redirections; out-of-stock soft 404s.
• Élevé impact / élevé effort — plan and schedule: faceted-nav architecture; Core Web Vitals fonctionner; rolling Product schema à travers templates; crawl-depth / architecture fixes.
• Low impact / low effort — quand temps permet: meta descriptions on long-tail products; minor title-template tidy-ups.
• Low impact / élevé effort — skip: redirect-chain cleanup on zero-traffic pages; Ouvrir Graph tags (social, pas ranking).

The whole point of the matrix is permission to not do choses.

Courant myths to clair up in the report

• “More indexed pages = better SEO.” Aucun — Google’s propre advice is to “eliminate contenu dupliqué to focus exploration on unique content plutôt que unique URLs.” Inflating the index with thin variant pages usually hurts.
• “Use noindex to save crawl budget on facets.” Aucun: “don’t utiliser noindex, as Google va encore requête, but alors drop lune page… wasting exploration temps.” Block the récupérer (robots.txt) or utiliser fragment filtering à la place.
• “Shopify handles all canonicalization.” It sets the balise canonical, but it doesn’t fix the internal-linking-to-non-canonical problem. Voir Step 3.
• “Crawl budget affects every store.” It doesn’t — la plupart stores jamais besoin to think à propos de it.

AI summary

A condensed prendre on the Avancé version:

• An ecommerce SEO audit trouve the few problèmes holding a store back — it’s a prioritized liste, pas a 500-point checklist. The philosophy is client-first (anchor to the réel pain) and “parfois the meilleur course of action is to do nothing.”
• Commencer dans la recherche Google Console and explorer données, pas on-page. On a store, the biggest wins are exploration/indexation problems at scale; “un mistake peut garder millions of pages out of the index.”
• Audit order: crawlability → indexation (GSC Page Indexation report) → duplicate content → on-page at scale → technical (CWV LCP<2,5s / INP<200ms / CLS<0,1, HTTPS, canonicals, mobile) → données structurées → lien internes → out-of-stock → off-page.
• #1 ecommerce problème: contenu dupliqué from faceted navigation, parameter permutations, and variants. Fix with robots.txt blocking or fragment filtering — not noindex. The Shopify catch: it canonicalizes /collections/.../products/... but lien internes encore point at the non-URL canonique (orphans).
• #2 problème: thin category pages. Mueller: hard to rank category pages que are simplement product grids — but a short utile block beats a giant article at the bottom.
• Données structurées (Product, Breadcrumb, Org, Examiner) rend vous eligible pour résultats enrichis; it doesn’t faire vous rank.
• Prioritize on impact/effort: rapide wins (stray noindex, dirty sitemaps, variant canonicals) premier; architecture (facets, CWV, schema rollout) planned; busywork skipped.
• Budget d’exploration is a non-issue pour la plupart stores — it matters at ~1M+ pages modification weekly or 10k+ daily, or with a big “Discovered – not indexed” bucket.

Documentation officielle

The principal sources behind chaque audit step.

Google — exploration & indexation

• Optimize votre budget d’exploration — who nécessite it, soft-404 waste, pourquoi pas to utiliser noindex to enregistrer budget.
• Managing exploration of faceted navigation URLs — l’URL-explosion problem and the robots.txt / fragment fixes.
• Page indexation report (Search Console Aider) — every “not indexed” status, notamment the duplicate-parameter remarque.
• Inspection d’URL outil (Search Console Aider) — checking a unique URL’s indexé state and données structurées.

Google — ecommerce specialty

• Ecommerce Structure d’URL meilleur practices — minimizing alternate URLs, descriptive paths, & separators, canonical parameter handling.
• Aider Google comprendre votre ecommerce site structure — the menu → category → sub-category → product hierarchy and <a href> requirement.
• Données structurées pour ecommerce sites — the recommended schema types.
• Product snippet données structurées — requis and recommended Product fields.

Google — technical

• Understanding Core Web Vitals and Recherche Google results — the LCP/INP/CLS thresholds and the CWV report.
• Résultats enrichis Tester — validate Product, Breadcrumb, and autre markup.

Bing / Microsoft

• bingbot Series: Maximizing Explorer Efficiency — Bing’s “crawl efficiency” framing; relevant to large-catalog stores.

Quotes from the source

On-the-record statements behind the audit. The deep liens jump to the quoted passage on the source page.

Me — audit methodology (from my Ahrefs writing)

• “If clients are coming to you asking for an audit, they already have a pain point. Talk to them. Solve that one thing and they’ll be happy with the audit.” — Patrick Stox, Free SEO Audit Template.
• “Anything high-impact and low-effort is a quick win, so those tasks should be tackled first.” — Patrick Stox, Enterprise SEO technique.
• “One mistake can keep millions of pages out of the index or remove an entire site from search results.” — Patrick Stox, Enterprise SEO technique.
• “Sometimes, the best course of action is to do nothing because the costs outweigh the benefits.” — Patrick Stox, We Studied Over 1 Million Domains.
• “Most sites don’t need to worry about crawl budget, but there are few cases where you may want to take a look.” — Patrick Stox, Quand Devrait Vous Worry À propos de Budget d’exploration?
• “Though it’s a joke in the SEO community, ‘it depends’ is really the answer when dealing with out-of-stock products on e-commerce websites.” … “Ultimately, there’s no perfect solution.” — Patrick Stox, How Devrait Vous Handle Out-of-Stock Products?.

John Mueller, Recherche Google Advocate — thin category pages (relayed)

• “When the ecommerce category pages don’t have any other content at all, other than links to the products, then it’s really hard for us to rank those pages.”
• “Maybe 90%, 95% of that text is unnecessary. But some amount of text is useful to have on a page so that we can understand what this page is about.”
• “Our algorithms sometimes get confused when they have a list of products on top and essentially a giant article on the bottom.”

Google — official docs

• “Eliminate duplicate content to focus crawling on unique content rather than unique URLs.” Jump to quote
• “Don’t use noindex, as Google will still request, but then drop the page when it sees a noindex meta tag or header in the HTTP response, wasting crawling time.” Jump to quote
• “The crawlers will typically access a very large number of faceted navigation URLs before the crawlers’ processes determine the URLs are in fact useless.” — Managing exploration of faceted navigation URLs.
• “Use <a href> tags when creating links to other content. Don’t use JavaScript events on other HTML DOM elements for navigation.” — Aider Google comprendre votre ecommerce site structure.

The ecommerce SEO audit checklist

Run top to bottom — the order is the prioritization. Don’t commencer at on-page.

1. Crawlability

• Lire robots.txt; nothing important disallowed, low-value parameter spaces blocked.
• “Blocked by robots.txt” report reviewed pour faux positives.
• GSC Statistiques d’exploration vérifié pour crawled-vs-indexed gaps (explorer waste).
• Tout URLs exported and clustered by parameter/chemin to size the faceted problem.
• Crawl-budget concern confirmed réel (1M+ weekly / 10k+ daily / big “Discovered”) avant acting.

2. Indexation

• GSC Page Indexation “Not indexed” exported by raison and quantified.
• Indexé count comparé contre connu catalog size.
• Inspection d’URL run on a sample from chaque raison bucket.
• XML sitemap contient seulement canonical, indexable, 200-status URLs.
• Aucun accidental sitewide noindex.

3. Contenu dupliqué

• Les deux GSC “Duplicate” buckets reviewed.
• Duplicate clusters with aucun canonical identified in the robot d’exploration.
• Faceted parameters handled via robots.txt or # fragments (pas noindex).
• Shopify /collections/*/products/* canonicalized — and the canonical is internally lié.
• HTTP/HTTPS, www/non-www, trailing-slash tout 301 to un canonical.

4. On-page at scale

• Manquant/duplicate titles, H1s, and meta descriptions filtered in the robot d’exploration.
• Thin category pages flagged; short utile copy block, pas a wall of text.
• Title templates vérifié pour collisions.

5. Technical

• CWV by URL groupe: LCP < 2,5s, INP < 200ms, CLS < 0,1.
• HTTPS everywhere; aucun mixed content.
• Canonicals: aucun 4XX targets, aucun non-canonical sitemap URLs, aucun paginated-to-page-one, unique tag par page.
• Mobile version isn’t stripped of description/schema/images.

6. Données structurées

• Résultats enrichis Tester passes on product + category templates.
• GSC Product Snippets / Merchant Listings reports clean.
• Merchant listing requis définir présent: name, image, offers (price/currency/availability).
• Product snippet eligibility présent: name plus au moins un of review, aggregateRating, or offers.
• Schema prices match visible prices; breadcrumb schema matches visible trail.

7. Maillage interne

• Orphan pages and explorer depth reviewed; clé pages dans ~3 clicks.
• Liens to redirections and broken liens pulled.
• Meilleur sellers lié from nav and hub/editorial pages.

8. Out-of-stock / discontinued

• Soft 404 bucket pulled and resolved by pattern.
• Temporary vs. permanent decisions applied (garder / 301 / 404-410).
• Lien internes to deleted products supprimé or mis à jour.

9. Off-page

• Branded vs. non-branded click share reviewed.
• Liens by page type assessed.
• 404s with backlinks 301’d to live pages.

10. Prioritize & deliver

• Every finding scored on impact/effort.
• Rapide wins premier; architecture planned; busywork skipped.
• Report focused on the few problèmes que matter, quantified in business impact.

The frameworks behind the audit

1. Scale-first ordering. The audit runs in priority order parce que leverage differs by area: an on-page tweak fixes un page; an indexation pattern fixes thousands. So: crawlability → indexation → contenu dupliqué → on-page → technical → données structurées → lien internes → out-of-stock → off-page. Resist the urge to ouvrir with title tags.

2. The impact/effort matrix. Score every finding on two axes and act by quadrant:

The matrix’s réel valeur is permission to skip — “parfois the meilleur course of action is to ne faites pashing.”

3. Client-first scoping. Don’t audit everything. Commencer from the pain the store en réalité has (trafic drop, a category que stopped ranking, a bad migration), solve que, alors widen. A focused report of 5–10 quantified problèmes beats a 200-row explorer export every temps.

4. The three “not equals” (inherited from SEO technique). Exploration ≠ indexation (a blocked page peut encore be indexé), exploration ≠ ranking (plus explorer ≠ plus élevé positions), exploration ≠ rendering (JS runs separately). La plupart ecommerce confusion — “why is my product not showing up?” — resolves une fois vous locate which stage it’s failing at.

5. Out-of-stock decision tree. Temporary → garder live (restock/waitlist). Permanent + fermer match → 301. Permanent + liens/trafic, aucun match → garder with connexe products. Permanent + nothing → 404/410. “It depends” is the honest par défaut; the variables are permanence and equity.

Outils pour an ecommerce SEO audit

Mentioned in ce audit

• Recherche Google Console — the backbone: Page Indexation report (every “pas indexé” raison), Statistiques d’exploration, Core Web Vitals, Sitemaps, Inspection d’URL, and Enhancements (Product Snippets, Merchant Listings). Free, and où the audit starts.
• Ahrefs Site Audit — my principal robot d’exploration: crawlability, indexability, duplicate clusters, canonicals, orphan pages, lien internes, explorer depth, performances, and structured-data checks à travers 170+ problème types.
• Ahrefs Site Explorer — backlink profile by page type, broken pages with liens (redirect-reclamation targets), organic keywords by page, competitor lien gaps.
• Google Résultats enrichis Tester — validate Product, BreadcrumbList, and autre schema on a live URL.
• PageSpeed Insights — field (CrUX) + lab CWV données on representative templates.

Free options

• Ahrefs Webmaster Outils — free explorer + Site Audit pour sites vous vérifier; the no-cost façon to obtenir la plupart of the ci-dessus.
• Bing Webmaster Outils — explorer info, Site Scan, and IndexNow (push price/stock changements au lieu de waiting pour a recrawl).

Autre robots d’exploration

• Screaming Frog SEO Spider — desktop bulk export of URLs, code d’états, canonicals, meta données, hreflang; handy pour very grand custom crawls.
• Chrome DevTools — confirmer si product content is in the initial HTML or injected by JavaScript (“View source” vs. “Inspect”).

Playbook: trafic organique drops à travers an ecommerce site

1. Confirmer the scope. Split Search Console données by page type, country, device, and requête class. Si seulement un template or market déplacé, garder the investigation là; si tout segments déplacé, continuer sitewide.
2. Align the timing. Mark deployments, migrations, feed changements, inventory events, seasonality, and connu search changements on the même timeline. Si the drop begins with a release, inspect que release avant compiling a generic problème liste.
3. Vérifier accès and réponse changements. Comparer current and prior robots rules, code d’états, canonicals, sitemaps, and rendered lien internes. Si important URLs are blocked, redirigé, noindexed, or orphaned, contain que échec premier.
4. Reconcile indexation. Join intended URLs from le sitemap/catalog with explorer results and Search Console Page Indexation raisons. Si the loss clusters in duplicate or canonical states, inspect URL signals; si it clusters in crawled-not-indexed, examine content and page valeur.
5. Tester representative templates. Validate un known-good and un affected category, product, facet, and editorial page pour rendering, schema, maillage interne, and Core Web Vitals. Si a shared component fails, widen the sample avant modification individual pages.
6. Prioritize by affected valeur and confidence. Ship the smallest reversible fix pour the highest-value confirmed causer. Si evidence is inconclusive, gather a targeted sample plutôt que bundling speculative changements.
7. Vérifier and annotate. Record the exact modifier, tester its technical output, and monitor the affected segment contre its propre baseline. Si the intended signal did pas modifier, roll back or reopen the diagnosis.

Audit practices que waste the team’s attention

Export every robot d’exploration warning and appel it an audit

Pourquoi it fails: severity étiquettes ne faites pas know le site’s templates, trafic, revenue, or intended URL policy. Do à la place: connecter chaque finding to affected URLs, evidence, business impact, and a spécifique fix owner.

Commencer with title-tag rewrites during an indexation loss

Pourquoi it fails: on-page polish ne peut pas fix blocked, redirigé, canonicalized, or unrendered pages. Do à la place: establish crawlability and indexation scope avant moving bas the stack.

Treat every excluded URL as a problem

Pourquoi it fails: alternate variants, filtered URLs, and noncanonical duplicates may be intentionally excluded. Do à la place: comparer the observed state with the documented indexation policy pour chaque URL class.

Modifier several systems avant measuring anything

Pourquoi it fails: combined template, canonical, content, and linking changements destroy causal clarity and complicate rollback. Do à la place: groupe connexe fixes, state the attendu signal, and validate chaque release.

Diagnose audit evidence que ne fait pas line up

Utiliser the shared SEO audit evidence package pour capture, timestamp, environment, scope, limitations, reproduction, and acceptance tests. Extend it ici with product ID, SKU, selected variant, market, postcode, inventory/fulfillment state, feed source and processing temps. Garder the original catalog, page, and feed captures separate from the analyst’s diagnosis.

Explorer totals are far ci-dessus the catalog size

Probable causer: faceted, sorting, tracking, search, or session parameters are generating URL combinations. Fix: classify the parameter patterns, inspect internal discovery and canonical behavior, and explorer a bounded sample avant recommending contrôle.

Search Console and the robot d’exploration disagree on indexability

Probable causer: the explorer sees today’s réponse pendant que Search Console reflects an précédent explorer, or rendering changements directives après raw HTML. Fix: comparer timestamps, raw and rendered output, canonical signals, and representative Inspection d’URL results.

Structured-data errors apparaître on seulement some products

Probable causer: optional catalog fields, variant logic, or out-of-stock states enter a différent template branch. Fix: segment errors by template and données condition, reproduce un affected record, and correct the shared mapping plutôt que hand-editing URLs.

Recommendations garder growing but nothing ships

Probable causer: findings lack impact, ownership, dependencies, or acceptance tests. Fix: convert chaque confirmed problème into a ticket with affected scope, evidence, proposed modifier, attendu signal, owner, and validation step.

Prompts pour organizing audit evidence

Cluster findings sans inventing severity

Paste a sanitized explorer/problèmes export après ce prompt.

Group these ecommerce SEO findings by root cause and affected template. Preserve the
original evidence and URL counts. For each group, return: observed signal, likely
system owner, evidence still needed, affected page type, reversible first test, and
validation method. Do not assign business impact or severity unless the input
contains traffic, revenue, or indexation evidence supporting it.

[PASTE AUDIT EXPORT]

Turn confirmed findings into implementation tickets

Convert only the confirmed findings below into engineering-ready tickets. Each ticket
must include current behavior, intended behavior, affected URL pattern, reproduction
steps, proposed acceptance tests, monitoring window, and rollback condition. Separate
facts from hypotheses and place unresolved questions in a final section.

[PASTE CONFIRMED FINDINGS AND EVIDENCE]

Ecommerce audit signal map

Mesurer si the audit program improves le site

Confirmed high-impact findings resolved

Metric: count and share of evidence-backed priority findings shipped and validated, pas merely closed. Ce que it indique vous: si audit fonctionner reaches production and produces the attendu technical signal. How to pull it: join the audit ledger with issue-tracker status and acceptance-test evidence. Benchmark / realistic range: establish a baseline by team capacity and dependency class; ne faites pas reward closure of low-value findings to inflate the percentage. Cadence: every sprint and quarterly by root causer.

Intended index coverage by page type

Metric: intended URL canoniques represented in the attendu indexation state, segmented by product, category, editorial, and approved facet pages. Ce que it indique vous: si the searchable inventory matches policy. How to pull it: reconcile sitemap/catalog URLs, explorer states, and Search Console Page Indexation exports. Benchmark / realistic range: the target dépend on explicit URL policy; excluded variants ne doit pas be counted as échecs. Cadence: monthly and après platform releases.

Organic performances of affected cohorts

Metric: clicks, impressions, and qualified organic outcomes pour the exact URL/requête cohorts tied to shipped fixes. Ce que it indique vous: si technically validated fonctionner corresponds with durable search and business improvement. How to pull it: enregistrer pre-change Search Console cohorts and join les to analytics or commerce outcomes où governance permet. Benchmark / realistic range: comparer chaque cohort with its propre seasonal baseline and an unaffected comparison groupe quand possible. Cadence: annotate at release, alors examiner après sufficient recrawling and monthly thereafter.

Recurrence rate

Metric: validated problèmes que reappear on the même template or URL class après remediation. Ce que it indique vous: si the root causer was fixed or seulement the current symptoms were patched. How to pull it: comparer scheduled explorer detectors and audit ledger fingerprints à travers runs. Benchmark / realistic range: utiliser the premier two comparable audits to définir a baseline and treat repeated systemic defects as prevention fonctionner. Cadence: chaque audit cycle.

Testez vos connaissances: ecommerce SEO audits

Five questions on audit sequencing, evidence, and prioritization.

Journal des modifications