Technical SEO Services: The 23-Point Audit Every Agency Should Run First
Most technical audits we inherit from previous agencies share one flaw: they list 200 "issues" with no thresholds, so nobody knows what's actually broken versus cosmetic. A page with a Largest...
Technical SEO Services: The 23-Point Audit Every Agency Should Run First
Short answer: Technical SEO services fix the crawlability, indexing, speed, and structured-data problems that stop search engines from ranking your pages. Before any keyword or content work, a competent provider runs a technical SEO audit — ideally a fixed checklist with clear pass/fail thresholds — so you fix what's actually blocking traffic, not what's easy to bill for.
Most technical audits we inherit from previous agencies share one flaw: they list 200 "issues" with no thresholds, so nobody knows what's actually broken versus cosmetic. A page with a Largest Contentful Paint of 2.6s and one at 9s both show up as "needs improvement" in the same red row. That's not an audit. That's a screenshot of a crawler. The fix is a scored checklist where every line has a number you either clear or you don't.
What are technical SEO services, exactly?
Technical SEO services are the work that makes a site crawlable, indexable, fast, and machine-readable — separate from content (what you say) and off-page (who links to you). Think of it as the plumbing. You can write the best page on the internet, but if Googlebot can't render it, your canonical points to a dead URL, or your schema throws errors, that page competes with one hand tied behind its back.
In 2026 the scope widened. It's no longer just "can Google crawl this." It's also: can ChatGPT, Perplexity, and Google's AI Overviews extract and cite your content. According to Google's own guidance to marketers, reiterated this month, generative engine optimization is still SEO — the same technical fundamentals that earn blue-link rankings make a page eligible for AI citation. The plumbing got more important, not less.
A real technical SEO engagement typically covers five pillars: crawlability, indexing, performance, structured data, and rendering — plus, increasingly, AI-crawler access. The audit is where it starts.
Why does the audit have to come first?
Run the work in the wrong order and you waste the retainer. Crawlability problems mask indexing problems: if Googlebot can't reach a page, auditing its canonical tags tells you nothing. Performance fixes on pages that are noindex by accident move zero rankings. We've seen six-figure content programs stall for a year because nobody checked that the blog template shipped with a sitewide noindex after a redesign.
The sequence that works:
- Crawlability — can bots reach the pages? (robots.txt, sitemap, internal links)
- Indexing — are the right pages in the index, and only those? (canonicals, noindex, duplicates)
- Performance — do real users get a fast page? (Core Web Vitals from field data)
- Architecture — is the site shallow, clean, and HTTPS-consistent?
- Structured data & rendering — can machines parse it? (schema, JS rendering)
- AI-search readiness — can generative engines extract and cite it?
Fix in that order and each layer's results are trustworthy. Jump around and you're guessing.
The 23-point technical SEO audit (with pass/fail thresholds)
Here's the part competitors leave out: a copyable checklist where every item has a threshold you either clear or you don't. This is the exact frame we run on a new growth-stage site before we touch anything else. Copy it, run it against your own domain, and you'll know in an afternoon whether your current provider is earning their retainer.
| # | Audit item | Pillar | Pass threshold | Fail trigger |
|---|---|---|---|---|
| 1 | robots.txt config | Crawl | No important pages, CSS, or JS disallowed | Money pages or render assets blocked |
| 2 | XML sitemap hygiene | Crawl | Only 200-status, indexable, canonical URLs | Contains 404s, redirects, or noindex URLs |
| 3 | Crawl budget waste | Crawl | <10% of crawls hit non-indexable URLs | Bots burning budget on params/facets |
| 4 | Internal link reach | Crawl | Every indexable page ≥1 internal link | Orphan pages with zero inbound links |
| 5 | Indexation gap (GSC) | Index | Submitted-vs-indexed gap <10% | Large "Discovered, not indexed" pile |
| 6 | Canonical tags | Index | Every page self-canonical or points to a live 200 | Canonical to redirect, 404, or wrong page |
| 7 | Accidental noindex | Index | No revenue page carries `noindex` | Any money page set to noindex |
| 8 | Duplicate / thin pages | Index | <5% near-duplicate clusters | Templated thin pages indexed at scale |
| 9 | LCP (field data) | Perf | <2.5s at 75th percentile mobile | ≥4s, or "Poor" in CrUX |
| 10 | INP (field data) | Perf | <200ms at 75th percentile | ≥500ms |
| 11 | CLS (field data) | Perf | <0.1 at 75th percentile | ≥0.25 |
| 12 | TTFB | Perf | <800ms | ≥1.8s |
| 13 | Mobile usability | Perf | Zero mobile errors in GSC | Tap targets / viewport errors |
| 14 | Click depth | Arch | Key pages ≤3 clicks from home | Revenue pages 4+ clicks deep |
| 15 | Redirect chains | Arch | No chain longer than 1 hop, no loops | Chains ≥3 hops or loops present |
| 16 | Broken links (4xx/5xx) | Arch | Zero broken internal links | Any internal 404/5xx in nav or body |
| 17 | HTTPS & mixed content | Arch | Full HTTPS, valid cert, no mixed content | HTTP assets or expired cert |
| 18 | Schema validity | Schema | 0 errors in Rich Results Test on templates | Errors on a templated page type |
| 19 | JS rendering | Render | Primary content in rendered HTML | Content only appears after client JS |
| 20 | Security headers | Render | `X-Frame-Options` set correctly | Missing — flagged by Google as SEO-relevant |
| 21 | AI-crawler access | AI | GPTBot / PerplexityBot not blocked unintentionally | AI bots disallowed in robots.txt by default |
| 22 | AI Overview eligibility | AI | Answer-first blocks + clean schema present | No extractable answer formatting |
| 23 | Entity & answer formatting | AI | Key facts in tables, lists, definitions | Facts buried in unstructured prose |
Score it 23/23. Anything under 20 and your content program is running uphill. Items 9–11 use the exact thresholds Google publishes for Core Web Vitals — LCP under 2.5s, INP under 200ms, CLS under 0.1, all measured at the 75th percentile of real Chrome users, not a one-off lab test that flatters your hardware. Item 20 reflects John Mueller's recent note that X-Frame-Options is the one security header genuinely tied to SEO.
You can run items 21–23 automatically with SEO Magics' AI Overview Checker, or get the full 23-point pass against your live URL through our free SEO audit tool. The point of the checklist isn't the tool — it's that you stop accepting "issues found: 184" as a deliverable.
How much do technical SEO services cost?
Pricing splits by whether you're buying a one-time audit or ongoing work. The ranges below come from Backlinko's 2026 SEO pricing data and industry pricing surveys, cross-checked against what we see quoted to growth-stage clients.
| Service | Typical 2026 range | What moves the price |
|---|---|---|
| One-time technical audit | $500 – $10,000 | Site size; automated report vs. manual analysis |
| Core Web Vitals optimization | $1,500 – $15,000 | Framework complexity, current performance gap |
| Schema markup implementation | $500 – $5,000 | Number of templated page types |
| Site architecture restructure | $3,000 – $25,000 | URL count, redirect mapping scope |
| Site migration support | $5,000 – $50,000+ | Platform change, traffic at risk |
| Monthly retainer (incl. technical) | From $2,500/mo | Scope, reporting cadence, output volume |
A useful filter: a $750 "audit" is almost always an automated crawl export with no prioritization. A genuinely useful audit costs more because a human reads the crawl, assigns severity, and hands you a sequenced fix list. If you want the deeper breakdown of models and company sizes, we cover it in How Much Do SEO Services Cost in 2026.
How long does technical SEO take to work?
Faster than content, slower than you'd like. Crawl and indexation fixes — unblocking robots.txt, repairing canonicals, cleaning a sitemap — often show up in Google Search Console within 2 to 8 weeks as pages get recrawled. Ranking and traffic gains usually take 3 to 6 months to fully materialize, and a large migration can need 6 to 12 months to stabilize.
The trap most founders fall into: expecting technical fixes to move rankings overnight. They don't, directly. What they do is remove the ceiling. A page that was capped at position 30 because it loaded in 9 seconds can finally compete once it loads in 2 — but it still has to earn the position with relevance and links. Technical SEO is permission to rank, not the ranking itself.
What does an ongoing technical SEO service include?
A one-time audit is a snapshot. Sites regress: a developer ships a release that reintroduces a redirect chain, a CMS update strips schema, a new section launches orphaned. Ongoing technical SEO is the monitoring layer that catches drift before it costs traffic.
A real ongoing service covers recurring crawls, Core Web Vitals monitoring from field data, schema validation after each release, log-file analysis for crawl-budget waste, and a standing fix queue prioritized by traffic impact. This is where the AI-native angle matters: we run the 23-point check on a schedule and flag the moment an AI Overview stops citing a page it used to cite. For growth-stage teams, that's usually packaged inside an SEO consultant retainer rather than billed as a separate technical line item.
Technical SEO vs. content SEO: which moves the needle?
Both, but in sequence. The honest answer most agencies won't give: if your technical foundation passes the 23-point check, you should spend the next dollar on content and links, not on more technical work. Technical SEO has a ceiling — once the plumbing works, polishing TTFB from 600ms to 500ms moves nothing.
Where technical wins outright is when something is genuinely broken: a noindex on money pages, a JS framework that hides content from crawlers, a migration that dropped 40% of URLs. Those are emergencies. Absent an emergency, technical is the gate you pass through once, then maintain — and content is the compounding engine. For where the two should actually overlap, our take on schema and AI-search rewards is the closest thing to a bridge between the two disciplines.
Methodology
The 23-point checklist in this article is the intake audit SEO Magics runs before onboarding a growth-stage client — refined across audits of SaaS, DTC, and ecommerce sites on retainer. Crawl and architecture items (1–8, 14–17) are validated with Screaming Frog and Google Search Console's Pages and Crawl Stats reports. Performance items (9–13) use field data from the Chrome User Experience Report rather than lab scores, against Google's published Core Web Vitals thresholds. Structured-data and rendering items (18–20) are checked with Google's Rich Results Test and the URL Inspection tool's rendered-HTML view. AI-search items (21–23) reflect our generative-engine optimization work and current guidance from Google and Search Engine Journal on AI citation. Pricing ranges are cross-referenced with published 2026 industry data; we cite ranges, not invented case-study numbers. The checklist is a framework, not a guarantee — every site has edge cases a fixed list won't catch, which is why a human reads the crawl before we assign severity.
Frequently asked questions
What is the difference between a technical SEO audit and a regular SEO audit?
A technical audit focuses on crawlability, indexing, speed, and structured data — the machine-readable foundation. A regular or full SEO audit adds content quality, keyword targeting, and backlink analysis on top. The technical audit comes first because content and link work assume the foundation is sound.
Can I run a technical SEO audit myself?
Partly. Free tools like Google Search Console, the Rich Results Test, and a crawler will surface most of the 23 items. What you usually can't do without experience is prioritize — knowing that a single orphaned template matters more than 50 cosmetic warnings. Run the checklist yourself; bring in help to sequence the fixes.
How often should a technical SEO audit be run?
A full audit every 3 to 6 months, plus a lighter regression check after any major change — a redesign, migration, CMS update, or core Google update. Sites drift, and the cost of catching a regression late is measured in lost rankings.
Do technical SEO fixes help with AI Overviews and ChatGPT citations?
Yes. Clean schema, fast rendering, crawlable HTML, and answer-first formatting are the same signals that make a page eligible for AI citation. You can check a specific page with the AI Overview Checker. Generative engines can't cite what they can't parse.
Is technical SEO a one-time project or ongoing?
The initial fix is a project; keeping the score at 23/23 is ongoing. Most regressions are introduced by routine development work, so the monitoring matters as much as the original audit.
How do I know if a technical SEO agency is any good?
Ask for their audit framework before you sign. If they can't show you thresholds — what counts as pass versus fail — they're selling crawl exports, not analysis. A good provider hands you a sequenced, severity-ranked fix list, not a 200-row spreadsheet.
Run your 23-point audit
Score your own site against the checklist above. If you land under 20, you have technical debt that's quietly capping every page you publish. Drop your URL into the free SEO audit tool for an automated first pass, or read more technical breakdowns in the SEO Magics journal. When you want a human to read the crawl and tell you exactly what to fix first — not a list of 184 "issues" — book a strategy call. We'll run the 23 points live and show you where the ceiling is.