Most technical SEO advice is a list of fifty items with no sense of which matter. In practice a handful of problems account for nearly all real damage, and the rest are tidiness.

Here is the list ordered by how much harm each one does.

Tier one: things that stop you ranking at all

Accidental noindex

The single most damaging and the most common. A <meta name="robots" content="noindex"> left over from staging, or applied to a template, removes the page from search entirely.

Check: Search Console → Pages → "Excluded by 'noindex' tag". Also view source on your key templates and search for "noindex".

robots.txt blocking

A disallow rule covering a directory you need indexed. Common after a migration where the staging robots.txt is deployed to production.

Check: visit /robots.txt and read it. Test key URLs in Search Console's robots.txt tester.

Worth noting: robots.txt prevents crawling, not indexing. A blocked page can still appear in results without a description, which looks worse than either alternative.

Pages returning the wrong status

A page serving 404 or 500 cannot rank. After a redesign this is the usual cause of traffic collapse, because old URLs with accumulated ranking history now return nothing.

Check: Search Console → Pages → "Not found (404)". Every URL there with historic traffic needs a 301.

Canonical pointing somewhere wrong

A canonical tag tells search engines which URL is authoritative. Pointing every page at the homepage — which happens with misconfigured templates — effectively deindexes the site.

Check: view source on several different pages and confirm each canonical points at itself.

Tier two: things that waste what you have

Redirect chains

URL A redirects to B redirects to C. Each hop loses a little, and crawlers give up after a few. Always redirect directly to the final destination.

Check: curl -sSIL <url> | grep -i location and count the hops.

302 where 301 belongs

A 302 is temporary. For a permanent move — which a redesign is — use 301, which passes ranking signals properly.

Incomplete sitemap

A sitemap listing a fraction of your pages tells search engines you do not consider the rest important. It should be generated from the same data that generates the pages, so it cannot drift.

Check: count the URLs in sitemap.xml against the real page count.

Orphan pages

Pages with no internal links pointing to them. They may be in the sitemap and still be treated as unimportant, because internal linking is how crawlers understand what matters.

Check: crawl the site and compare discovered URLs against the sitemap.

Duplicate content from parameters

Filters, sorting and tracking parameters creating many URLs with the same content. Handle with canonicals pointing at the clean version.

Tier three: things that limit how well you do

Core Web Vitals

Largest Contentful Paint under 2.5 seconds, Interaction to Next Paint under 200 milliseconds, Cumulative Layout Shift under 0.1. A modest ranking factor and a significant conversion factor.

Check: Search Console → Core Web Vitals, using field data rather than lab scores.

Mobile parity

Indexing is mobile-first, so the mobile version is the version that counts. Content hidden on mobile is content Google may not see.

Check: compare mobile and desktop rendering of your key templates, and test with a real phone rather than a resized browser window.

Structured data

Not a direct ranking factor, but it governs rich results and how reliably AI systems can extract your content. Worth having, correctly:

  • Organization with sameAs links on the site
  • BreadcrumbList on every deep page
  • Article on blog posts, with author and dates
  • FAQPage wherever you have genuine questions and answers
  • LocalBusiness only where you actually have a presence
  • Product and Offer on ecommerce

Check: Google's Rich Results Test and Schema Markup Validator. Search Console reports errors under Enhancements.

Heading structure

One <h1> per page, headings in order, no skipping levels, and headings used for structure rather than styling. Matters for accessibility, for crawlers and increasingly for AI extraction.

HTTPS and host canonicalisation

One canonical host. Apex or www, not both answering independently, with a 301 from the other. Mixed content warnings resolved. SSL valid and auto-renewing.

Check: curl -sSI https://www.example.com | head -5 should show a 301 to your canonical host, or vice versa.

Internal linking

Important pages should be reachable within a few clicks of the homepage and linked from relevant context. A footer linking every service and location page is a cheap and effective way to keep crawl depth shallow.

Tier four: worth doing, rarely urgent

  • Image alt text that describes the image
  • Descriptive, readable URL slugs
  • hreflang, if you genuinely serve multiple languages
  • An XML sitemap index, once you exceed 50,000 URLs
  • Pagination handled sensibly on long archives
  • llms.txt for answer engines

How to run this

Do not work through it as a list. Start in Search Console:

  1. Pages report — anything excluded that should not be, and all 404s
  2. Core Web Vitals — which metric fails, on which templates
  3. Enhancements — structured data errors
  4. Performance — pages that lost impressions, which points at what broke

That surfaces the real problems in twenty minutes. Then crawl the site with a tool to catch what Search Console does not report — orphan pages, redirect chains, missing canonicals.

Most sites turn out to have three or four genuine problems, usually in tier one or two. Fixing those is worth more than working through everything else.

If you want this run properly on a specific site, with the problems ranked by estimated impact, get in touch.