Vorlagen
SEO Audit Checklist Template
A checklist in the order things block each other, so you never spend an afternoon on content for a page that turns out to be noindexed.
Starte ein frisches DomainLens-Audit und nutze den Report als Prioritätenliste.
How to use this
Work top to bottom. The sequence is dependency-ordered: each section assumes the ones above it passed, because a page that cannot be indexed gains nothing from better headings.
Audit a representative URL per template rather than every page. Findings almost always belong to a template, and checking twenty product pages tells you what checking one would have.
1. Indexability — nothing below matters until these pass
| Check | Pass looks like | Where |
|---|---|---|
| Status code | 200 for real pages | Response headers |
| robots.txt | Path not disallowed | /robots.txt |
| Meta robots | No unintended noindex | HTML head |
| X-Robots-Tag | No unintended noindex | Response headers |
| Canonical | Self-referencing, resolves 200 | HTML head |
| Google-selected canonical | Matches your declaration | URL Inspection |
| Rendering | Content present without JavaScript | curl the URL |
2. Structure and discovery
| Check | Pass looks like |
|---|---|
| Sitemap | Exists, declared in robots.txt, parses, 200 URLs only |
| Internal links in | The page is linked from somewhere relevant |
| Internal links out | Real anchors with href attributes |
| Click depth | Reachable within three or four clicks |
| Redirect chains | One hop maximum |
| Broken links | None on the page |
| Hostname | One canonical host; the other redirects |
3. On-page
| Check | Pass looks like |
|---|---|
| Intent match | Format matches what ranks for the query |
| Title | Unique, specific, front-loaded |
| Meta description | Written for the click, not duplicated |
| H1 | One, stating the subject |
| Heading outline | Levels descend without skipping |
| Content depth | Answers the question completely |
| Images | Alt text, dimensions, sensible format |
| Structured data | Valid, and matches visible content |
4. Performance and delivery
Field data first — a lab score is a diagnostic, not the target. See Core Web Vitals vs Lighthouse score .
| Check | Pass looks like |
|---|---|
| LCP | Under 2.5s in field data |
| INP | Under 200ms in field data |
| CLS | Under 0.1 in field data |
| TTFB | Under 800ms |
| Compression | Text responses served gzip or brotli |
| Cache-Control | Long on fingerprinted assets, short on HTML |
| HTTPS | Valid certificate, no mixed content |
Why the order is dependency, not severity
Most audit checklists are grouped by topic — one section for technical, one for content, one for links. That reads tidily and produces a bad working order, because it lets someone spend a morning rewriting titles on pages that turn out to be excluded from the index.
Dependency ordering costs nothing and prevents that entirely. Each section here assumes the ones above it passed, so a failure stops you and sends you back rather than letting you continue into work that cannot pay off.
It also changes what a partial audit is worth. Getting through sections one and two and stopping still leaves you with a usable result, because you have established that the pages can be found and indexed. Getting through the content section without checking indexability leaves you with findings that may all be moot.
Recording what you find
For turning this into an ordered plan see how to prioritise SEO fixes , and for presenting it see the audit report example .
- 1Note the check, the URLs affected, and how many share the cause.
- 2Group by cause immediately — forty findings from one template is one item.
- 3Mark each as blocking, structural, content or polish.
- 4Assign an owner while the finding is fresh, or it will not get one.
- 5Write down what you decided not to fix, and why.
- How often should I run this?
- A full pass once or twice a year, with continuous monitoring between. Running it monthly usually means re-reporting the same unfixed findings.
- Do I check every page?
- No. One representative URL per template. Findings belong to templates far more often than to individual pages.
- What if a check does not apply?
- Mark it not applicable and move on. A checklist that forces irrelevant findings trains people to ignore it.