Why technical SEO becomes urgent at scale

A five-page brochure site can get away with sloppy technical SEO. A ten-thousand-SKU catalog with color, size, and filter combinations cannot — search engines allocate a finite crawl budget to your site, and a technically inefficient large catalog wastes most of it on low-value or duplicate URLs instead of your money pages. This checklist is organized by the most common failure points at scale.

Crawlability and indexation

  • [ ] robots.txt doesn't accidentally block important sections. A misconfigured disallow rule (common after a platform migration) can silently deindex an entire category.
  • [ ] XML sitemap is current, submitted, and error-free in Google Search Console — check the "Pages" report for a rising count of "Discovered, not indexed" or "Crawled, not indexed" pages, which signals a quality or duplication problem, not just a crawling one.
  • [ ] Faceted navigation (filters) doesn't generate infinite unique URLs. Filter combinations (color + size + price range, in any order) can produce a combinatorial explosion of near-duplicate, low-value URLs. Use canonical tags, noindex on deep filter combinations, or robots.txt rules to control this.
  • [ ] Out-of-stock and discontinued product pages have a defined policy — either kept live with a clear "back in stock" path and internal links to alternatives (usually best for popular items expected to return), or 301-redirected to the closest replacement/category (for permanently discontinued items) rather than left as unhandled 404s or, worse, silently removed with no redirect.
  • [ ] Pagination is implemented cleanly — each paginated category page (page 2, 3, etc.) should be a distinct, indexable URL rather than infinite-scroll content with no unique URL per "page."

Canonicalization

  • [ ] Every page has exactly one canonical URL, and variant/filtered/sorted versions of a page point their canonical tag back to the primary version.
  • [ ] No conflicting signals — a page shouldn't have a canonical tag pointing to URL A while also being included in the sitemap under URL B.
  • [ ] HTTPS and www/non-www consistency — pick one version and 301-redirect all others to it; mixed signals here split ranking authority across duplicate versions of the same page.

Site speed and Core Web Vitals

  • [ ] Largest Contentful Paint (LCP) — the time for the main content (often the hero product image) to render; commonly slow on ecommerce sites due to unoptimized, oversized images.
  • [ ] Cumulative Layout Shift (CLS) — unexpected layout jumps as a page loads (a common ecommerce culprit: ads, reviews widgets, or promotional banners that load in without reserved space and push content around).
  • [ ] Interaction responsiveness — how quickly the page responds to a shopper's first interaction (tapping "Add to Cart," opening a size selector); slow third-party scripts (chat widgets, some review platforms, marketing pixels) are common culprits.
  • [ ] Image optimization — compressed, appropriately sized, and served in modern formats (e.g., WebP) with lazy-loading below the fold.

Structured data at scale

  • [ ] Product, Review, and BreadcrumbList schema implemented at the template level, not manually per page — verify a sample across product types with Google's Rich Results Test.
  • [ ] Structured data stays accurate as prices and stock change — schema that's static or cached separately from the live page can drift out of sync with what's actually displayed, which risks a manual action if flagged.

International and multi-currency considerations (if applicable)

  • [ ] hreflang tags correctly implemented if you serve multiple country or language versions of the same catalog, so search engines show the right regional version to the right searcher.
  • [ ] Duplicate content across regional sites is managed with hreflang and/or canonicalization rather than left to compete against itself.

A prioritization framework, since you can't fix everything at once

When facing a long list of technical issues, prioritize roughly in this order: (1) issues actively blocking indexation of important pages, (2) large-scale duplicate-content problems (faceted navigation, out-of-stock handling), (3) Core Web Vitals failures on your highest-traffic templates, (4) structured data gaps or errors, (5) smaller cosmetic issues. A single robots.txt error blocking your whole category structure matters more than a missing alt tag on a handful of images.

Common mistakes

  • Discovering a robots.txt or noindex misconfiguration only after organic traffic has already dropped noticeably, rather than through routine monitoring.
  • Letting faceted navigation generate thousands of indexed low-value URLs, diluting crawl budget away from real category and product pages.
  • Fixing Core Web Vitals on a staging environment but never verifying the fix shipped correctly to the live site's real-world (field) data.
  • Treating a technical SEO audit as a one-time project rather than a recurring check, especially after platform migrations or major theme updates.

Best practices

  • Check Google Search Console's Pages report on a regular cadence (at minimum monthly for a growing catalog) rather than only when traffic visibly drops.
  • Set a firm, written policy for out-of-stock and discontinued product URLs before you need one at scale.
  • Re-run a Core Web Vitals check after any theme, app, or major content change — third-party scripts are the most common regression source.
  • Keep an explicit canonicalization strategy documented so anyone editing templates knows the rule.

Troubleshooting

  • "Crawled, not indexed" count is rising in Search Console — usually a content-quality or duplication signal, not a crawl-access problem; check for thin or near-duplicate pages first.
  • A whole category disappeared from search seemingly overnight — check robots.txt and any recently deployed noindex rules first; this is the most common cause of a sudden, sitewide drop.
  • Core Web Vitals fail in the field but pass in lab testing (e.g., PageSpeed Insights lab data) — field data reflects real users on real devices/connections; check for third-party scripts, ad networks, or apps that behave differently under real-world network conditions.

FAQs

How often should a large catalog run a full technical SEO audit? At least quarterly, plus immediately after any platform migration, major theme change, or a noticeable unexplained traffic drop.

Is faceted navigation inherently bad for SEO? No — it's a UX necessity for large catalogs. The risk is unmanaged faceted navigation generating unlimited unique, indexable URLs. Managed well (canonical tags, selective noindex, or robots rules on deep combinations), it's not a problem.

Do Core Web Vitals actually affect ranking, or just user experience? Both — they're a documented ranking factor, though a relatively modest one compared to content relevance and quality; they matter more as a conversion-rate factor than most sellers initially assume.