Most SEO audits fail for a structural reason rather than a technical one. The findings are correct, the analysis is sound, and nothing gets implemented — because the document is organised the way the auditor investigated rather than the way the reader has to act.
This is the template we use. Copy the structure; the value is in the ordering and in one column most audits omit.
The template
| Section | What goes in it | Length |
|---|---|---|
| 1. Verdict | The single sentence answer, plus the three things to do first | ½ page |
| 2. Priority actions | Every finding worth fixing, ranked, with effort and owner | 1–2 pages |
| 3. Crawl & indexation | Can Google reach and store the pages that matter | 1–2 pages |
| 4. On-page & content | Whether pages deserve the rankings they target | 2–3 pages |
| 5. Site architecture & internal links | How authority moves through the site | 1–2 pages |
| 6. Off-site | Link profile, brand signals, anything toxic | 1 page |
| 7. Appendix | Full URL lists, crawl exports, raw data | Unlimited |
Seven sections. The order is not negotiable, and the reason is section 1.
Why the verdict comes first
The person commissioning an audit usually is not the person implementing it. They need to make a decision — approve budget, brief a developer, fire an agency — and they will read the first page carefully and skim the rest.
An audit that opens with methodology and works toward conclusions forces that reader to do the synthesis themselves. Most will not. They will forward the PDF to someone technical, who will read section 3, fix two crawl issues, and consider the audit delivered.
Open with the answer:
Verdict: Organic traffic has fallen 34% since March because 1,100 product pages were canonicalised to category pages during the June template change. Nothing is wrong with the content or the link profile. Fix the canonical logic and expect recovery within two crawl cycles.
Do these three first: (1) Remove the hardcoded canonical from the product template. (2) Resubmit the product sitemap. (3) Stop the weekly content spend until indexation recovers — it is not the constraint.
That is the entire audit for the decision-maker. Everything after it is evidence.
The column most audits leave out
Section 2 is a table, and it needs four columns, not three:
| Finding | Impact | Effort | Owner |
|---|---|---|---|
| Product canonicals point to categories | High — 1,100 pages deindexed | 2 hrs (dev) | Engineering |
| No H1 on 340 category pages | Medium | 4 hrs (CMS) | Content team |
| 12 orphaned blog posts | Low | 1 hr | Content team |
| Missing Organization schema | Low | 30 min (dev) | Engineering |
Effort is the column that gets things fixed. Audits that rank findings by impact alone produce a list where item one costs three months of engineering time, so nothing starts. Add effort and the reader immediately sees that finding four is thirty minutes and can be done today. Momentum matters more than sequence.
Rank by impact, but never hide the effort. If you cannot estimate effort in hours, you do not understand the fix well enough to recommend it.
Section 3: crawl and indexation
This section answers one question: can Google reach, render and store the pages that are supposed to earn money?
Work through the crawlability, indexation and site-structure items in our technical SEO checklist — robots.txt, sitemap validity, crawl errors, canonical logic, noindex usage, redirect chains. Do not reproduce the checklist in the audit. Report only what is broken, with counts and example URLs.
The pattern to watch for: a page count mismatch. If the CMS says 4,000 pages, the sitemap says 3,200, and Search Console reports 1,900 indexed, the gap is the finding. Most large traffic losses live here rather than in content quality. Our guide to diagnosing a traffic drop walks the same tree in more detail.
Section 4: on-page and content
Assess whether the pages deserve to rank for what they target, not whether they follow a rule.
- Intent match — does the page format match what Google currently returns for the query? A 3,000-word essay targeting a query that returns tools and templates will not win, however well written.
- Cannibalisation — two pages competing for the same query. Check whether both earn impressions for the same terms in Search Console.
- Thin and duplicate — pages that exist for structural reasons and satisfy nobody.
- Titles and meta — worth a paragraph, not a section. See title tags and meta descriptions for the specifics rather than expanding them here.
Resist the per-page table. A spreadsheet of 400 rows with red and green cells looks thorough and gets ignored. Group into patterns: “All 340 category pages share one template flaw” is one finding, not 340.
Section 5: architecture and internal links
The most under-served section in most audits and often the cheapest wins, because it needs no new content and no developer.
Report click depth from the homepage, orphaned pages with no internal links pointing at them, and whether your commercially important pages actually receive internal links from relevant articles. Our internal linking guide covers the method; the audit reports only the gaps.
The finding usually writes itself: “31 of your 49 articles link to no commercial page.” That is a sentence a client can act on this week.
Section 6: off-site
Keep it short unless something is actually wrong. Referring domain trend, anchor text distribution, and any pattern suggesting a manual action rather than an algorithmic decline.
Do not pad this section with a competitor backlink gap analysis unless the client asked for one. It belongs in a link-building proposal, not an audit.
Section 7: appendix
Every URL list, crawl export and screenshot goes here. Nothing in sections 1–6 should be longer than a page because the supporting data lives at the back.
This is also what separates an audit from a sales document. A real audit ships its raw data so the client can verify the findings — and take them to another supplier if they want to. If a supplier will not release the crawl export, ask why.
Three things to cut
The competitor overview. Three pages describing what competitors do well, with no action attached. Interesting, not useful.
Tool score screenshots. A screenshot of a “SEO score: 62/100” dial is not a finding. The underlying issues are the findings.
Everything the client already knows. If they told you in the kickoff that the blog has not been updated in two years, do not present it back as a discovery.
Template versus worked example
This page gives you the empty structure. Filling it well is a separate skill — the difference between “missing meta descriptions on 340 pages” and a finding someone can act on comes down to how it is written. We have three annotated findings, written badly and then rewritten, in SEO audit report samples.
If you need the recurring monthly version rather than a one-off diagnostic, that is a different document with different sections — see the SEO report template.
When to get one done for you
You can run this yourself with a crawler and a Search Console login, and for a site under a few hundred pages that is usually the right call.
Get help when traffic has dropped and you have already looked without finding the cause, when the site is large enough that the patterns are not visible by hand, or when you need a second opinion on work someone else did.
Which service depends on what prompted it. If traffic has already fallen, our traffic drop audit is the narrower, faster and cheaper answer. If nothing has broken and the question is whether the foundation can support the rankings you are aiming at, that is a technical SEO audit. Both use exactly this structure and ship the appendix data with it. If you are weighing that against a broader engagement, the pricing comparison sets out what each option costs — and what an audit should cost covers the audit specifically.
