The SEO audit checklist for non-experts (the 4-question method)

Most SEO audits fail at the last step: they get read, nodded at, and filed. Four hundred warnings sorted by category is a report; it isn't a plan. The audits that produce rankings share one property — every finding answers four questions:

1. What's wrong? (specifically, on which pages)
2. Why does it matter? (what it costs in plain terms)
3. How do I fix it? (the actual change, not "optimize your metadata")
4. Did it improve? (a re-check, later, of the same thing)

We built our whole product around those four questions, but you can run the method manually. Here's the checklist, ordered by payoff-per-hour.

Round 1: Can Google even see the site? (30 minutes)

• robots.txt isn't blocking anything important — test it, especially if traffic ever "mysteriously" dropped after a launch.
• No accidental noindex on money pages — check headers and meta together with a page check.
• The sitemap exists, parses, and matches reality — validate it, then compare its URL count with what you actually publish.
• HTTPS works and the certificate isn't near expiry — 30 seconds to check.
• One canonical host — apex vs www, one 301s to the other, every page self-canonicalizes on the winner.

Round 1 problems are rare but fatal; everything else is pointless while one exists. If a page you care about isn't in Google, run the indexing decision tree before touching anything cosmetic.

Round 2: Does each important page state its case? (an afternoon)

For your ten most important pages — not all pages, the ten that matter:

• Unique, specific title in the 50–60 character budget, query language first.
• Meta description that pitches the click, 150–160 characters.
• Exactly one H1 that agrees with the title.
• Alt text on images that carry meaning.
• Internal links in — at least three, from relevant pages, with descriptive anchors.
• Schema — breadcrumbs everywhere, plus the type that matches the page.

A free page check runs all of these at once and scores the result, which makes the before/after in Round 4 measurable instead of vibes.

Round 3: The site-level problems you can't see per-page (needs a crawl)

Some issues only exist in aggregate: duplicate titles across pages, orphan pages with zero internal links, redirect chains, pages three-plus clicks deep. These need something that fetches your whole site and compares pages against each other — which is precisely the job of a crawler-based audit. This is the round where a tool stops being optional; manual spot checks structurally cannot find comparison problems.

Round 4: Did it improve? (the round everyone skips)

Whatever you fixed, put a date on it and re-measure:

• Re-run the same page checks — did the scores move?
• In Search Console, watch the affected pages' queries for two to four weeks — position and CTR, not just clicks.
• Keep the list of what you changed. When something improves, you want to know which change did it; when nothing improves, that's the signal to stop polishing details and address content or links.

An audit without Round 4 is a mood. With it, it's a loop — and the loop, run monthly, is the entire practical difference between sites that grind upward and sites that stay put.