Inventory what is indexed
Open the index coverage report and export reality. Sort by patterns: filters, sort parameters, session IDs, thin tag pages, printer views, staging leftovers.
Then put your money URLs beside that list. The uncomfortable gap between “what we sell” and “what is indexed” is the brief for the next sprint.
If you cannot explain why a URL is indexed, it is a candidate for consolidation or removal.
Canonicals and duplicates
Conflicting canonicals are silent killers. Self-canonical to A, sitemap says B, internal links prefer C, hreflang points at a fourth host. Crawlers pick something. It may not be the page you rewrote last month.
Pick one preferred URL per document. Make links, sitemaps, and headers agree. Parameter variants should not compete for the same intent unless you have a deliberate reason and the content truly differs.
Templates that tax every URL
A slow product template multiplied by ten thousand URLs is not a performance ticket. It is an SEO ticket. So is a render path that hides primary content behind client-side calls the crawler never waits for.
Fix the shared template before you commission another round of blog posts. New content on a broken chassis underperforms and then gets blamed on “the algorithm”.
Fix or noindex with intent
Every wasted indexed URL should get a decision: improve it until it deserves a crawl, redirect it to a stronger URL, or noindex it and stop linking to it.
Do not noindex as a panic button for pages that should rank. Do not keep linking to noindexed URLs from the main navigation either. Intent goes both ways.
Technical SEO done well is housekeeping with a ledger. When the ledger is clean, content and links finally have somewhere solid to land.