TL;DR: An SEO audit verifies six things in dependency order: indexation, crawl access, on-page tags, content rendering, page speed, and links. Five independent datasets covering more than 400,000 websites show the same 30 to 60 percent failure rates on these checks. Free tools cover the whole process in two to four hours.
Most SEO audit guides give you a flat list of 50 to 100 items and leave the order to you. The order is the hard part.
Fixing titles on a page Google cannot crawl wastes the afternoon. Tuning page speed on a URL blocked by robots.txt changes nothing.
We run audits for clients in the USA, UK, and beyond, and we run the same process on our own site at clienvora.com. This guide is that process, written out in full, with the failure rates each check carries across five independent industry datasets.
| Audit depth | Time | Cost | What you get |
|---|---|---|---|
| This guide, free tools only | 2-4 hours | $0 | Six dependency-ordered checks, fix list ranked by severity and effort |
| Paid crawler (Screaming Frog, Sitebulb) | 1-2 hours | $259+/yr | Same six checks with deeper crawl data and log analysis |
| Agency audit (like ours) | 1-3 weeks | $1,500+ | Technical, content, and competitive layers with a prioritized roadmap |
If you have never audited a site before, the free path below will find more problems than you expect. Start there. Pay for depth only after the free checks stop finding new issues.
What an SEO Audit Actually Checks
An SEO audit checks whether search engines can find, crawl, understand, and rank your site. It runs in a fixed dependency order.
It is not a PageSpeed score. It is not a list of missing meta tags from a browser extension. A real audit answers one question at every layer: can Google reach this page, parse it, trust it, and serve it to the right query?
That breaks into six layers. Each layer only matters if the layer before it works.
Order beats volume. A site with 12 broken canonicals and a clean crawl path will outrank a site with perfect titles that Google cannot render. The six checks below run in that order on purpose.
The Six Checks to Run in Dependency Order
Run these six checks in order. Each lists the free tool, the pass criterion, and how often that check fails across the audited web.
| # | Check | Free tool | Pass criterion | Failure rate (source) |
|---|---|---|---|---|
| 1 | Indexation | Search Console Pages report | Every money page shows "Indexed" | 23% of sites have at least 1 URL blocked-but-indexed (Matcha Growth, n=312) |
| 2 | Crawl access | robots.txt tester + GSC URL Inspection | robots.txt returns 200, no stray Disallow on real pages | 87% of owners who want deindexing use the wrong fix (Matcha Growth) |
| 3 | On-page tags | Screaming Frog free (500 URLs) or Site Checker | Unique title, meta description, one H1, self-canonical per page | 31% of homepages have no canonical tag (Greadme, n=308) |
| 4 | Content rendering | URL Inspection "View crawled HTML" vs live DOM | Rendered HTML matches raw HTML for key content | 31.4% of sites show raw-vs-rendered mismatch (SEOmator, 100K sites) |
| 5 | Speed and Core Web Vitals | PageSpeed Insights + GSC CWV report | LCP under 2.5s, INP under 200ms, CLS under 0.1 at 75th percentile | 58.4% fail mobile LCP; 29.6% fail INP (SEOmator); 91% fail at least 1 CWV (Crawlix) |
| 6 | Links and trust | GSC Links report + Site Checker broken-link scan | Zero 4xx internal links; HTTPS everywhere; no redirect chains | 36.1% have broken internal links (median 14 per site); 24.3% have redirect chains 2+ hops (SEOmator) |
Two numbers in that table deserve a second look. The crawl-access failure rate is not about crawlers failing to reach your site. It is about owners making the wrong fix when they try to remove pages, which we cover in the mistakes section.
While you are in robots.txt, test Bingbot separately in Bing Webmaster Tools' Robots.txt Tester. The two crawlers interpret wildcard rules differently, and what passes for Googlebot does not always pass for Bingbot. Check that you are not blocking GPTBot, ClaudeBot, or PerplexityBot unless you intend to be invisible to AI search.
The speed row is the most over-invested check in SEO. More sites fail indexation and on-page basics than fail page speed. Fix speed after the earlier layers are clean, unless the GSC Core Web Vitals report already flags your money pages.
Start at check one. Skip nothing. Order matters.
What the Largest Audit Datasets Actually Find
Five independent audit datasets covering more than 400,000 websites agree on which issues are most common, and they disagree on one item worth naming.
| Issue | SEOmator (100K sites) | Greadme (643 sites) | SlapMyWeb (263 sites) | SE Ranking (418K sites) |
|---|---|---|---|---|
| Missing/duplicate meta description | 52.4% | Only 34.8% of homepages get all six on-page basics right (Greadme, n=305) | ||
| Images missing alt text | 51.8% (#3 issue) | n/a | 58% | 74% (#1 issue) |
| No structured data on eligible pages | 47.9% | n/a | 67% | n/a |
| Missing HSTS header | 38.6% | 0 of 112 passed all 5 security headers | 39% | n/a |
| Broken internal links | 36.1% (median 14/site) | n/a | n/a | n/a |
| Duplicate titles | 33.7% | n/a | n/a | n/a |
| Mobile LCP over 2.5s | 58.4% (#1 issue) | 69% imply lab LCP over 4s (n=207) | ~36% fail 2.5s LCP for real users | n/a |
| No canonical tag | n/a | 31% of homepages (n=308) | n/a | n/a |
One honest disagreement sits inside that table. SEOmator ranks missing alt text third at 51.8 percent. SE Ranking ranks it first at 74 percent across 418,125 audits.
Both numbers can be true because the populations differ. SEOmator's panel is roughly 60 percent B2B SaaS. SE Ranking's audit base skews ecommerce, and image-heavy stores fail alt-text checks more often.
A 52 percent meta-description failure rate does not mean your site has a 52 percent chance of failing. It means roughly half of audited sites fail that check. Treat the rates as a map of where audits usually find blood, then measure your own site against the pass criteria instead of the averages.
The Greadme dataset adds a correlation that most checklists hide. Across 625 paired scores, the correlation between performance and meta-tag quality was 0.00.
A fast site tells you nothing about its SEO basics. The two layers fail independently. That is another reason to run checks in order instead of trusting one dashboard score.
Two on-page checks run inside the same Search Console session. First, run your top 20 URLs through Google's Rich Results Test and fix schema errors. Ahrefs found schema errors on 37 percent of pages, most often missing required fields or using Article schema on non-article pages.
Second, filter the Performance report to a target query. If two or more of your pages appear, you have cannibalization. Google is splitting signals between them, and neither ranks as well as a single consolidated page would.
Only 6.8 percent of SEOmator's 100,000-plus sites scored an A. The median score was 71 out of 100. Audit findings are the normal condition of the web, not a sign that something unusual is wrong with your site.
Most sites fail something. Yours is not special here. Read the table twice.
How to Prioritize Fixes: Severity Times Effort
Rank every finding by traffic at risk and fix effort. Clear the high-traffic, low-effort items first. The matrix below is that ranking.
| Tier | Typical findings | Expected impact | Fix window |
|---|---|---|---|
| This week (high severity, low effort) | Money pages showing "Crawled, not indexed"; stray Disallow in robots.txt; broken canonical on a key URL; redirect chains on paid landing pages | Recovers pages Google already knows about | 1-3 hours |
| This month (high severity, medium effort) | Duplicate titles and meta descriptions on templates; missing H1s; 4xx internal links; raw-vs-rendered content gaps on JS pages | Improves snippet eligibility across templates | 1-2 weeks of part-time work |
| This quarter (medium severity, high effort) | Mobile LCP over 2.5s on top templates; INP failures; structured data rollout; HSTS and security headers | Strengthens ranking signals and rich results | 1-2 sprints with a developer |
Crawlix puts a number on the ceiling. Shipping the ten most common fixes typically yields 10 to 25 percent organic traffic improvement over 60 to 90 days. That range assumes the earlier layers were broken in the first place.
Do not start in the bottom tier because it feels more technical. PageSpeed work is the most satisfying audit finding to fix and often the least impactful one to fix first.
Rank first. Fix next. Re-measure after.
A Worked Example: Auditing Clienvora.com
We ran this exact six-check process on clienvora.com in October 2026, and the findings are published here with their limits stated.
Honest limits first. This is one site, self-audited by its own team, so the sample size is n=1 and the bias is ours. What follows still shows what a real small-business audit output looks like when nothing is staged for marketing.
Check one found a history problem more than a current one. In 2026 our build pipeline rewrote CSS media queries into CSS Range Syntax, and 73 range conditions across 41 files silently killed all mobile styles in older browsers.
Googlebot's mobile-first crawler saw the broken layout for weeks before we caught it in a PageSpeed run. Check rendering on real mobile devices, not just the crawler.
Check two surfaced our biggest structural debt. The site carries more than 450 rules in a single _redirects file accumulated since the Blogger migration.
Most are single-hop and harmless. A handful form chains, and those chains sit on URLs we still link to from older posts.
Check three was the cleanest layer. Titles and meta descriptions are unique on every template. The gap we found was stale internal links pointing at five renamed blog slugs, which we fixed in the same week.
Nothing in this section is a benchmark. One site, one industry, one migration history. What the example does prove is that a free-tools audit of a small business site surfaces fixable issues within two hours, which is the claim this guide started with.
Check five (speed) confirmed the mobile LCP issue above once we measured real-user data through the Chrome UX Report instead of lab tests alone. Check six (links) found the GSC linking-pages sample mostly consists of properties we control ourselves, which is exactly the kind of honest reading that vanity dashboards hide.
What Google Says About Auditing and Indexing
Google staff say three things about indexing. You cannot force pages into the index. Most pages on most sites never get indexed.
Search Console reports should guide your fixes, not your panic. Treat the coverage report as a change detector.
John Mueller put the first point plainly in June 2021.
"You can't force pages to be indexed. It's normal that we don't index all pages on all websites. It's not an issue with that page, it's more site-wide. Creating a good site structure and making sure the site is of the highest quality possible is essentially the direction."
In a July 2022 office-hours session he went further on scale. "Googlebot will never index everything across a website. I don't think it happens to any kind of non-trivial sized website that Google would go off and index completely everything." He added that "you really want Googlebot to focus on the important pages."
The Search Off the Record podcast turned that into operating advice.
"Don't think of it as an inventory that you need to fix. Don't think that everything on the website needs to be indexed. Don't think that the ratio between indexed and non-indexed is some sort of quality measure or some sort of score that you need to look at. Look at it as a detection tool for changes or patterns in the site that you expected or didn't expect."
Two practical rules follow from those quotes. First, stop treating a low indexed-to-submitted ratio as a failure state; treat it as a change detector. Second, below a million pages, spend crawl-budget anxiety on something else, like the broken canonicals in check three.
Google's own indexing documentation closes the loop on timing: "Some changes might take effect in a few hours. It may take considerably longer for Google to recrawl and reindex a site." Recrawl requests are not instant, which is why the fix window in our priority matrix is measured in weeks, not minutes.
Fixes are not instant. Weeks are normal. Patience is a tactic.
How Often to Run an SEO Audit
Run a full six-check audit quarterly, a five-minute Search Console check monthly, and an immediate re-check after any redesign, migration, or traffic drop.
The quarterly cadence matches how fast site debt accumulates in practice. Template changes, new plugins, CMS updates, and content publishes each introduce crawlable errors on a timeline of weeks. A quarterly sweep catches them before a Google core update makes them expensive.
| Cadence | Scope | Time | Who does it |
|---|---|---|---|
| Weekly (5 min) | Search Console coverage report, error count by type | 5 minutes | Anyone on the team |
| Monthly (30 min) | Core Web Vitals report, top-page indexation status | 30 minutes | Marketing lead |
| Quarterly (2-4 hours) | All six checks, full fix list, prioritized roadmap | 2-4 hours | SEO owner or agency |
| Event-driven | Re-run checks 1-4 after any migration, redesign, or platform update | 1-2 hours | Whoever shipped the change |
Event-driven re-audits matter most. Our own CSS media-range bug ran in production for weeks because no re-check ran after a build-pipeline change.
The quarterly cadence would have caught it later. The event-driven rule catches it the same day.
Quarterly full. Monthly light. Weekly glance.
When to Hire an Auditor Instead of Doing It Yourself
Hire an auditor when the site passes a million pages, revenue depends on organic traffic, or nobody has fixed crawl bugs before.
Below that threshold, the free path in this guide finds the same first-order issues an agency finds in week one. Our own SEO audit services check 200-plus items across seven layers because client sites carry CMS debt, faceted navigation, and migration history that a two-hour pass cannot reach.
The line to watch is not site size. It is fix ownership. Ship it in-house or buy help.
An audit produces a ranked list. If nobody on your team can edit templates, robots.txt, and server headers, the list rots. Hire implementation help with the audit, or buy the audit with implementation bundled.
For what a paid technical audit should include before you spend money, our guide to what to expect from a technical SEO audit lists the deliverables and the red flags. For how audit findings roll into a wider program, what to expect from an SEO agency in the first six months documents the timeline.
Five Mistakes That Waste the Whole Audit
Most wasted audits share five mistakes: wrong fix, skipped validation, chasing tool scores, treating site: queries as inventory, and never re-running the crawl.
1. Disallow-only deindexing
87 percent of owners who wanted pages removed from Google tried a Disallow rule in robots.txt alone. That blocks crawling, not indexing, and any page already indexed stays indexed.
The correct move is allow crawling plus a noindex tag, then removal in Search Console. Only 6 percent of audited owners got this sequence right (Matcha Growth, n=312).
2. Never clicking Validate Fix
After you fix a coverage issue, Search Console waits for you to press Validate Fix. Only 12 percent of audited owners ever clicked it.
Pressing the button triggers a faster recrawl of the affected URLs. Skipping it can leave correct fixes sitting in the "failed" column for weeks.
3. Chasing a PageSpeed score of 100
The lab score is not a ranking factor. The field metrics are.
Google's thresholds are LCP under 2.5 seconds, INP under 200 milliseconds, and CLS under 0.1, measured at the 75th percentile of real users. Only 42 percent of all websites pass all three (HTTP Archive / CrUX, 2026).
Passing the field thresholds beats scoring 100 in a lab run on a throttled connection.
4. Treating site: as an index inventory
The site: operator returns an approximate sample, not a complete index. As the Search Off the Record episode put it, do not treat the ratio between indexed and non-indexed pages as a quality score. Use the Search Console Pages report for counts and the site: operator only for spot checks.
5. Fixing pages before templates
Fixing one product page while the product template still ships duplicate titles means the same bug returns with the next import. Fix the template, then sweep the pages it generates. Templates are how 52.4 percent meta-description failures spread across a whole site instead of one URL.
6. Never re-running the crawl
An audit is a loop, not an event. Without a re-run, you cannot tell a fix that worked from a fix that half-worked, and new issues ship every sprint. The quarterly re-run is what turns the audit into a system.
One more quiet mistake: auditing everything equally. Not every finding is worth the same afternoon.
Rank, then fix, then re-measure. Then, and only then, widen the scope.
For the layer most sites skip entirely, our free SEO content audit tool grades a page against the on-page half of this checklist in under a minute, and the content grader does the same for the writing itself.
Frequently Asked Questions About SEO Audits
These seven answers cover audit timing, free tools, fix order, cadence, and when a paid audit beats a DIY pass.
How long does an SEO audit take?
A free-tools audit of a small business site takes two to four hours if you follow the six checks in order. A paid crawler cuts that to one to two hours by automating the crawl. A professional agency audit usually runs one to three weeks, because it adds content quality, competitive gaps, and a prioritized roadmap for someone else to implement.
What tools do I need for an SEO audit?
Four free tools cover the six checks. Google Search Console covers indexation, crawl errors, and Core Web Vitals. PageSpeed Insights covers lab and field speed. The robots.txt tester covers crawl access. Screaming Frog's free tier (500 URLs) or a broken-link checker covers tags, redirects, and internal 4xx errors. Paid crawlers add depth, not access.
What should I fix first after an SEO audit?
Fix indexation and crawl-access problems first, because everything downstream depends on Google reaching and indexing the page. That means money pages stuck as "Crawled, not indexed," stray Disallow rules, and broken canonicals. Then move to on-page tags on templates, then rendering, then speed, then links. Skipping ahead wastes work on pages Google cannot serve.
Can I do a full SEO audit for free?
Yes, for a site under about 500 pages. Google Search Console, PageSpeed Insights, the robots.txt tester, and Screaming Frog's free tier find the same first-order issues a paid tool finds on a small site. Paid tools earn their cost when you need full-site crawls, log-file analysis, scheduled re-runs, or historical comparison across thousands of URLs.
How often should an SEO audit be redone?
Full six-check audit: quarterly. Search Console error check: weekly, five minutes. Core Web Vitals and top-page indexation: monthly. Re-run checks one through four immediately after any redesign, migration, CMS update, or plugin change. Event-driven re-audits catch build-pipeline bugs that quarterly cadence alone would leave in production for weeks.
What is the difference between a technical SEO audit and a content audit?
A technical SEO audit covers crawl, index, render, speed, and link integrity: the layers that decide whether Google can serve a page at all. A content audit covers whether the page deserves to rank: intent match, depth, uniqueness, internal links, and conversion path. A full audit runs both, and the technical layer always comes first because good writing on an unindexable page produces nothing.
Do I need an SEO audit if my traffic is growing?
Yes, on a lighter cadence. Growing traffic hides accumulating debt, because old wins cover new mistakes until a core update removes the cushion. The datasets show a median audit score of 71 out of 100 across 100,000 sites, so the average site carries fixable issues while it grows. A quarterly light pass keeps the gap from compounding.
Do one thing today. Open Google Search Console, go to the Pages report, and filter for "Crawled - currently not indexed." Export the list. Those are pages Google already knows about and chose not to index, which makes them the fastest audit win on any site.
The technical half of this checklist is what we run for clients under our SEO audit services, which cover 200-plus items across seven layers. For the checklists behind each layer, start with the technical SEO audit checklist, the on-page audit checklist, the Core Web Vitals checklist, and the crawlability checklist.
About the Author's Work
Review the portfolio and the delivery process before reaching out, so you can evaluate the output before evaluating the person.
Get an SEO Audit Client Portfolio Free SEO Content Audit Start a Conversation How Clienvora WorksThe same standards appear in the technical SEO services, in how we read Google index coverage issues, and in the Core Web Vitals thresholds we audit against. Once the fix list exists, measure what moved with SEO KPIs and metrics.
(Disclosure: author's own service)
