Every technical SEO audit produces a long list. The skill is knowing which five items on that list are suppressing visibility right now, and which forty are noise that can wait. This is the prioritization framework we use on real client sites.
Indexation problems come first — always
If Google can’t index a page, nothing else you do to it matters. Start in Search Console’s Page Indexing report: pages marked "Crawled — currently not indexed" or "Discovered — currently not indexed" represent content Google found and rejected. On most sites we audit, a meaningful share of commercially valuable pages sits in these buckets without anyone noticing.
The common causes rank roughly in this order: thin or duplicative content, weak internal linking that signals low importance, canonical tags pointing somewhere unexpected, and crawl budget spent on parameter URLs and filter combinations instead of real pages. Fix indexation before touching anything cosmetic.
Redirect chains and broken canonicals leak equity quietly
A single 301 passes almost all authority. A chain of three or four — accumulated through years of migrations and URL cleanups — loses measurable equity at every hop and slows crawling. Crawl your site and export every redirect path longer than one step; flattening chains is boring work with disproportionate payoff.
Canonicals deserve the same suspicion. We regularly find templates that canonicalize paginated pages to page one (hiding deep products from crawl paths), HTTPS pages canonicalizing to HTTP ghosts, and self-referencing canonicals that break when URL parameters appear. Every canonical should be an explicit decision, not a template accident.
Severity is a function of the money the affected pages are supposed to make — not the color a tool assigns to the warning.
Core Web Vitals matter — after the crawl basics
Page experience is a real ranking signal, but it’s a tiebreaker between relevant results — not a substitute for indexable, relevant content. Chasing a perfect Lighthouse score while category pages sit unindexed is the most common misallocation of technical effort we see.
When you do optimize, use field data (the Chrome UX Report) rather than lab scores, and fix in this order: LCP through image and server response optimization, INP through JavaScript reduction, CLS through reserved space for images, embeds, and ads. Field thresholds — LCP under 2.5s, INP under 200ms, CLS under 0.1 — are the targets that count.
Structured data: high leverage, low effort, frequently broken
Schema markup doesn’t directly improve rankings, but it controls how your results look — prices, ratings, FAQs, breadcrumbs — and eligibility for rich results that lift click-through rates. It also feeds the entity understanding that AI-generated answers increasingly rely on.
Audit what’s deployed before adding more: validation errors in Product, Review, or Organization markup are extremely common after theme updates, and invalid schema is worse than none. Then extend to the types your content genuinely supports — never mark up content that isn’t visible on the page.
Build the priority list by revenue exposure, not error count
Tools sort findings by severity labels that know nothing about your business. Re-sort them yourself: which template does each issue touch, how much revenue flows through that template, and does the issue block indexing, dilute relevance, or merely offend best practice?
A missing meta description on a legal page is a non-issue. The same gap across 4,000 product pages is a template fix worth shipping this sprint. Severity is a function of the money the affected pages are supposed to make.
The useful takeaway
Fix what blocks indexing, then what leaks equity, then what slows the experience — in that order, weighted by the commercial value of the affected templates. A short list executed beats a long list admired.
