Fix anything preventing Google or AI crawlers from reaching your content. Unintended noindex tags. Robots.txt blocking key sections. Soft 404s on revenue pages. Broken canonicals. Blocked AI crawlers. Usually one-line config changes — but the difference between invisible and crawlable.
GEO
By Jelena Misić
·
Apr 27, 2026
·
28 min read
Olibro Resource · Pillar 02
Technical SEO Audit Checklist for 2026
The 147-point technical SEO audit framework Olibro runs on every client site. Organized by ROI so you fix the things that move rankings first.
By Jelena Misić
·
Apr 27, 2026
·
28 min read
TL;DR
A technical audit, in two minutes.
What's covered
Nine foundational layers — crawlability, indexation, Core Web Vitals, JavaScript rendering, structured data, mobile-first, international, security, AI crawler readiness — plus the table-stakes tenth layer most audits skip.
How it's organized
Each of the 147 checks is tagged Critical (ship in 7 days), Important (ship in 30), or Enhancement (quarterly). Most sites we see are missing 40–60 of these.
2026 shifts that change the playbook
Google's December 2025 Rendering Update can quietly exclude JavaScript pages from the index. AI crawlers — GPTBot, PerplexityBot, ClaudeBot — drive meaningful brand visibility, and most don't run JS.
How to use this page
Skim the explorer, expand a layer, see what your site already passes. Bookmark and revisit quarterly.
01 / Why 2026 is different
Three structural shifts that reframe what a technical audit even is.
If your last audit predates these, it's measuring the wrong things. We rebuilt our internal framework around them in Q1 2026.
i.
INP is a stable ranking signal
Interaction to Next Paint replaced FID in March 2024. By 2026 it's fully mature — Google weights it alongside LCP and CLS, and 200ms at the 75th percentile of mobile traffic is non-negotiable for competitive queries. Audits that still mention FID are outdated.
ii.
Google's December 2025 Rendering Update
Pages returning non-200 status codes — 404s, 5xxs, soft 404s — may be excluded from the rendering queue entirely. For JavaScript-heavy sites this is a silent killer: a misconfigured error page served with status 200 instead of 404 can cause Google to render junk and rank it over your real pages.
iii.
AI crawlers are a first-class concern
GPTBot, PerplexityBot, ClaudeBot, OAI-SearchBot, Google-Extended now drive meaningful brand visibility. Most don't execute JavaScript. If your hero, pricing, or docs require JS to render, you're invisible in ChatGPT, Perplexity, and Claude — regardless of how well you rank on Google.
02 / The 147-point framework
Every check we run. Ten layers, three priority tiers.
Tap any layer to expand or collapse the checklist. Most sites we see are missing 40–60 of these checks — start with Critical.
147 checks
Across 10 layers
54 Critical
Ship in 7 days
58 Important
Ship in 30 days
Critical
Important
Enhancement
01
Crawlability
7 8 3 +
Crawlability
If Googlebot and AI crawlers can't reach your content, nothing else matters. Most ranking issues in 2026 trace back to crawlability or indexation — audit these before optimizing anything else.
1
robots.txt is present and reachable at /robots.txt with a 200 status code
critical
curl, browser
2
robots.txt does not accidentally block Googlebot, Bingbot, or key sections
critical
GSC, robots.txt Tester
3
robots.txt does not block critical JS, CSS, or image files that affect rendering
critical
Screaming Frog, GSC URL Inspection
4
XML sitemap exists, is referenced in robots.txt, and submitted in GSC
critical
Google Search Console
5
XML sitemap contains only canonical, indexable URLs (no redirects, 404s, noindex)
important
Screaming Frog
6
XML sitemap is under 50MB uncompressed and under 50,000 URLs per file
important
File size check
7
Sitemap index file is used if multiple sitemaps exist
important
Manual review
8
No redirect chains longer than 2 hops on important URLs
important
Screaming Frog
9
No redirect loops anywhere on the site
critical
Screaming Frog
10
Internal links do not point to 4xx or 5xx URLs
important
Screaming Frog
11
No orphan pages (important URLs not linked internally)
important
Screaming Frog + GSC
12
Site architecture is 3 clicks or fewer from homepage to any key page
enhancement
Crawl depth
13
Breadcrumbs implemented on all interior pages with BreadcrumbList schema
enhancement
Rich Results Test
14
Pagination uses rel="next" only for user navigation, not as a canonical signal
enhancement
Manual review
15
GSC Crawl Stats show stable or rising pages crawled per day, not a sudden drop
critical
Google Search Console
16
Server response time (TTFB) is under 800ms on the 75th percentile
important
CrUX, PageSpeed Insights
17
No infinite crawl traps from faceted nav, session IDs, or URL parameters
critical
Log file analysis, GSC
18
URL parameters with no SEO value are blocked via robots.txt or canonicals
important
Manual + Screaming Frog
Critical 2026 flag
Check robots.txt today for accidental blocks of GPTBot, PerplexityBot, ClaudeBot, or Google-Extended. Some default CMS installs and "privacy-focused" plugins now block these by default. If you're not appearing in AI responses, this is the first place to look — not your content.
02
Indexation
6 7 3 +
Indexation
A crawled-but-not-indexed page is functionally identical to a page that doesn't exist. Google's index budget is tighter than it has been in a decade — your job is to signal that key pages meet the bar.
19
No unintended noindex tags on important pages (check X-Robots-Tag header too)
critical
Screaming Frog, GSC
20
Self-referencing canonical tags on every indexable page
important
Screaming Frog
21
No canonicals pointing to different domain versions (www vs non-www, http vs https)
critical
Manual review
22
No duplicate title tags across indexable pages
important
Screaming Frog, GSC
23
No duplicate meta descriptions across indexable pages
enhancement
Screaming Frog
24
GSC Page Indexing: zero pages with "Indexed, though blocked by robots.txt"
critical
Google Search Console
25
GSC: "Discovered – currently not indexed" is under 10% of crawlable URLs
important
Google Search Console
26
GSC: "Crawled – currently not indexed" investigated URL by URL for thin content
important
GSC + manual review
27
No soft 404s (pages returning 200 with "page not found" content)
critical
GSC Coverage report
28
Custom 404 page returns HTTP 404 status (not 200 or 302)
critical
curl, status checker
29
Pagination follows the "all content accessible via crawl path" rule
important
Manual review
30
Faceted nav URLs handled via canonicals, noindex, or robots.txt — one strategy, not mixed
important
Manual audit
31
Thin content pages (under 300 words with no unique value) noindexed or removed
important
Screaming Frog + review
32
Staging/dev/preview environments blocked from indexing (noindex + password + different domain)
critical
Manual verification
33
Printable, email, and PDF duplicates are canonicalized or noindexed
enhancement
Manual review
34
URL structure consistent: lowercase, hyphens not underscores, no session IDs in canonicals
enhancement
Manual review
The soft-404 trap
A "this product is no longer available" page returning HTTP 200 is a soft 404. Google detects the contradiction between status and content, demotes the page, and in the worst case loses trust in adjacent URLs. Return proper 404, or 301 to a relevant alternative — never 200.
03
Core Web Vitals & Performance
5 10 7 +
Core Web Vitals & Performance
LCP, INP, and CLS are confirmed ranking signals. Missing any one threshold at the 75th percentile of mobile traffic puts you in "needs improvement" — competitive queries now require passing all three.
35
LCP under 2.5s on 75th percentile mobile (CrUX field data, not lab)
critical
GSC CWV, CrUX
36
INP under 200ms on 75th percentile across all interactions
critical
GSC CWV
37
CLS under 0.1 on 75th percentile throughout the full session
critical
GSC CWV
38
TTFB under 800ms on 75th percentile
important
PSI, DebugBear
39
Hero images use modern formats (WebP or AVIF) with JPEG fallback
important
DevTools, Lighthouse
40
Images have explicit width and height attributes to reserve layout space
important
HTML source review
41
Above-the-fold images are NOT lazy-loaded (delays the LCP element)
critical
PageSpeed Insights
42
Below-the-fold images use loading="lazy" attribute
enhancement
HTML source review
43
Fonts preloaded with rel="preload" and font-display: swap
important
PageSpeed Insights
44
Render-blocking JavaScript minimized, deferred, or async where possible
important
Lighthouse
45
Unused JavaScript removed via code splitting or tree shaking
enhancement
DevTools Coverage
46
Third-party scripts audited quarterly — each one's performance cost justified
enhancement
WebPageTest
47
CSS minified and critical CSS inlined for above-the-fold content
important
PageSpeed Insights
48
Long tasks (>50ms) broken up using scheduler.yield() or requestIdleCallback
important
DevTools Performance
49
CDN in place for static assets
enhancement
Manual verification
50
Brotli compression enabled for text assets (better than gzip)
enhancement
curl -I
51
HTTP/2 or HTTP/3 enabled on the origin server
enhancement
HTTP/2 test tools
52
Browser caching headers correct (1 year for versioned assets, shorter for HTML)
enhancement
curl -I
53
Layout shifts from late-loading ads or embeds prevented via reserved space
important
PSI CLS debug
54
Web font fallback chain prevents FOIT (flash of invisible text)
important
Browser DevTools
55
Mobile CWV tracked separately from desktop — mobile is typically 30–50% worse
critical
GSC device filter
56
CWV improvements verified on the 75th percentile, not the median (Google uses p75)
important
CrUX, GSC
Our biggest INP win
An LA e-commerce client came to us with 1,400ms INP at p75. Refactored the product filter with debouncing and scheduler.yield(), moved third-party scripts to load after user interaction, cut category-page JavaScript by 40%. INP dropped to 180ms in two weeks. Mobile conversion rose 18%.
04
JavaScript Rendering
5 7 2 +
JavaScript Rendering
Google renders JavaScript with delays and limits. AI crawlers — GPTBot, PerplexityBot, ClaudeBot — mostly cannot render JS at all. If your content needs JS to appear in the DOM, you have a silent visibility problem across an increasingly important share of search surfaces.
57
Critical content (H1, copy, prices) renders in initial HTML, not only after JS
critical
View source, JS-off test
58
SSR or static generation used for public, crawlable pages
critical
Architecture review
59
Site tested with JavaScript disabled — key pages still render meaningfully
important
Browser DevTools
60
Key URLs fetched as Googlebot Smartphone in GSC URL Inspection — content matches
critical
GSC URL Inspection
61
Internal links are real anchor tags with href, not JS onClick handlers
critical
Source code review
62
Images have src attributes set in HTML, not only applied via JavaScript
important
Source view
63
Pagination uses crawlable links with href, not JS-only infinite scroll
important
Manual test
64
No JavaScript redirects (use HTTP 301/302 server-side instead)
critical
Screaming Frog
65
No content hidden behind clicks/hover that crawlers can't trigger
important
Source inspection
66
Large JS bundles code-split so initial render doesn't require the full bundle
enhancement
Bundle Analyzer
67
JavaScript console errors don't prevent page rendering on common browsers
important
DevTools console
68
Hydration mismatches between server and client render are logged and fixed
important
Framework dev mode
69
Next.js / Nuxt: ISR or static export used for catalog-style pages where possible
enhancement
Framework config
70
AI crawler rendering tested: fetch as GPTBot/PerplexityBot/OAI-SearchBot, confirm parity
important
curl with custom UA
The soft-404 JavaScript trap
Google's December 2025 Rendering Update means JavaScript sites that serve "friendly" error pages with HTTP 200 may have those pages excluded from the rendering queue entirely. If you have a React-based site that shows an error UI on 404 but returns 200, Google may never render real content and may index junk. Crawl with Screaming Frog and filter for pages returning 200 with "not found" in the title.
05
Structured Data
7 4 4 +
Structured Data
Schema.org JSON-LD has moved from "nice to have" to essential. Google uses it for rich results, Knowledge Graph, and AI Overview source selection. Content with proper schema gets 30–40% higher visibility in AI-generated answers across most verticals.
71
Organization schema sitewide with logo, sameAs (social profiles), and contact info
critical
Rich Results Test
72
Article schema on all blog posts, including author, datePublished, dateModified, publisher
critical
Rich Results Test
73
Author profiles link to real people with bios, sameAs to LinkedIn or academic profiles
important
Manual review
74
FAQ schema on pages with Q&A sections (not on pages without — Google penalizes misuse)
important
Rich Results Test
75
Product schema on product pages with price, availability, review rating where applicable
critical
Rich Results Test
76
LocalBusiness schema on contact pages with NAP, hours, geo coordinates
critical
Rich Results Test
77
BreadcrumbList schema on all interior pages (improves SERP and AI extraction)
important
Rich Results Test
78
Review/AggregateRating schema is real, not manipulated, and matches the page
critical
Policy + manual review
79
HowTo schema on tutorial content with step-by-step instructions
enhancement
Rich Results Test
80
All schema implemented as JSON-LD (not Microdata or RDFa, which are deprecated)
enhancement
Source view
81
Schema validated with Google's Rich Results Test — no errors
critical
Rich Results Test
82
Schema markup reflects visible on-page content (no hidden "schema-only" data)
critical
Manual audit
83
GSC Enhancements report: no structured data errors or warnings on key page types
important
Google Search Console
84
VideoObject schema on pages with embedded videos, including duration and thumbnail
enhancement
Rich Results Test
85
Event schema on event pages with startDate, endDate, location, offers
enhancement
Rich Results Test
The schema mistake that triggers manual actions
FAQ schema on pages without an FAQ section, or Review schema with manufactured ratings, is a policy violation Google increasingly enforces with manual actions. Implement schema only when the on-page content supports it — exactly as visible to users.
06
Mobile-First Compliance
5 4 4 +
Mobile-First Compliance
Mobile-first indexing has been Google's default since 2019 — meaning Google uses the mobile version as the primary version for crawling, indexing, and ranking. Mobile is not secondary; desktop is the afterthought.
86
Mobile and desktop versions contain identical primary content — no hidden critical text
critical
Side-by-side comparison
87
Mobile version has the same structured data as desktop
critical
Rich Results Test (both)
88
Mobile viewport meta tag correctly set (width=device-width, initial-scale=1)
critical
HTML source
89
Mobile navigation links are in the DOM, not only after hamburger click
critical
View source with JS off
90
Tap targets minimum 48×48 pixels with adequate spacing
important
Lighthouse, PSI
91
Font sizes readable without zooming (minimum 16px body text)
important
DevTools responsive
92
No horizontal scrolling on any viewport width from 320px up
important
Browser DevTools
93
Form inputs use correct type attributes (email, tel, number) for mobile keyboards
enhancement
Source review
94
autocomplete attributes set on form fields to enable browser auto-fill
enhancement
Source review
95
No intrusive interstitials on landing pages (confirmed ranking demotion signal)
important
Manual mobile test
96
Swipe-only carousels have button alternatives for accessibility
enhancement
Keyboard nav test
97
Hover-only interactions have tap equivalents for touch devices
enhancement
Mobile device test
98
No separate m. subdomain; legacy mobile sites 301 redirect 1:1 to responsive equivalents
critical
Screaming Frog (mobile UA)
07
International & Multilingual
4 6 1 +
International & Multilingual
For many LA businesses this layer applies even without international expansion — Los Angeles County is 55% non-English-speaking at home, and targeting Spanish, Korean, Chinese, or Persian audiences uses the same hreflang infrastructure.
99
hreflang tags are bidirectional (if A points to B, B must point back to A)
critical
hreflang Testing Tool
100
hreflang language-region codes follow ISO standards (es-US, not es-la or spanish)
critical
Manual review
101
x-default hreflang set for the fallback version (usually English global)
important
Manual review
102
Each language/region version is self-canonical (does not canonical to another language)
critical
Screaming Frog
103
URLs differentiate language clearly (/es/, /en/, or subdomain), consistent across the site
important
Architecture review
104
Translated content is professionally translated or transcreated, not machine translation only
important
Linguistic review
105
Language switcher uses hreflang-matching links, not JavaScript-only switching
important
Source view
106
Character encoding is UTF-8 for all pages (no Windows-1252 or Latin-1)
critical
curl -I, source
107
Right-to-left languages (Arabic, Persian, Hebrew) use proper dir="rtl" attribute
important
Source review
108
Currency, date format, and number format adapt to locale
enhancement
Manual review
109
GSC International Targeting report shows no hreflang errors
important
Google Search Console
08
Security & HTTPS
7 2 3 +
Security & HTTPS
HTTPS has been table stakes for years, but 2026 has added layers: HSTS preload, TLS 1.3 expectations, mixed content detection in Chrome, and stricter CSP. Security is also a ranking factor — recovery from a hack can cost months of traffic.
110
Site served over HTTPS with a valid, non-expired SSL certificate
critical
Browser, SSL Labs
111
All HTTP URLs 301 redirect to HTTPS equivalents
critical
Screaming Frog
112
No mixed content warnings (HTTP resources loaded on HTTPS pages)
critical
DevTools console
113
HSTS (Strict-Transport-Security) header set with max-age of 1+ year
important
curl -I, securityheaders.com
114
Site submitted to the HSTS preload list (for high-security sites)
enhancement
hstspreload.org
115
TLS 1.2 or higher enforced (TLS 1.0/1.1 are deprecated)
critical
Qualys SSL Labs
116
Qualys SSL Labs grade is A or A+
important
ssllabs.com
117
Content Security Policy (CSP) header in place (reporting at minimum)
enhancement
securityheaders.com
118
X-Content-Type-Options, X-Frame-Options, Referrer-Policy headers set
enhancement
securityheaders.com
119
Admin login pages have rate limiting and 2FA enabled
critical
Platform config review
120
Site not flagged by Google Safe Browsing (check GSC Security Issues)
critical
Google Search Console
121
CMS core, themes, and plugins updated — no known vulnerabilities in active components
critical
Patchstack, Wordfence
09
AI Crawler Readiness
5 6 5 +
AI Crawler Readiness
Most sites we audit have never considered AI crawler access deliberately — they're either accidentally blocking major crawlers or accidentally allowing none. This layer is also where the highest-ROI 2026 wins live, because competitors haven't optimized here yet.
122
GPTBot is allowed in robots.txt (unless deliberately blocked)
critical
robots.txt review
123
PerplexityBot is allowed in robots.txt
critical
robots.txt review
124
ClaudeBot is allowed in robots.txt
critical
robots.txt review
125
OAI-SearchBot (ChatGPT Search) is allowed in robots.txt
critical
robots.txt review
126
Google-Extended configured intentionally (allowed for Gemini, blocked if opting out)
important
robots.txt review
127
Applebot-Extended configured intentionally (for Apple Intelligence)
enhancement
robots.txt review
128
Critical content in initial HTML (AI crawlers don't execute JavaScript)
critical
curl with AI UA
129
Key pages verified by fetching as GPTBot user-agent — content parity
important
curl -A "GPTBot/1.2"
130
llms.txt at /llms.txt with key pages, titles, descriptions in markdown
enhancement
Manual creation
131
Content uses question-based H2s that mirror natural language queries
important
Content audit
132
Direct answers (40–60 words) appear under each H2 before expansion
important
Content audit
133
Every post has a TL;DR or summary at the top
important
Content audit
134
Statistics and factual claims include source attribution inline
important
Content audit
135
Author bios include credentials and sameAs links to LinkedIn / professional profiles
enhancement
Manual review
136
"Last updated" dates visible and current on evergreen content
enhancement
Template audit
137
Server logs monitored monthly for AI crawler traffic volume and behavior
enhancement
Log file analysis
10
Foundations
3 4 3 +
Foundations
The basics that don't fit cleanly into one layer but still matter — measurement, sharing, on-page SEO hygiene, and the meta-check that ensures this list gets run again next quarter.
138
Google Search Console verified for all domain and subdomain properties
critical
Google Search Console
139
Google Analytics 4 installed and measurement is working
critical
GA4 DebugView
140
Favicon present at multiple sizes for browser tabs, PWA icons, SERP
enhancement
Manual check
141
Open Graph and Twitter Card meta tags present and accurate for social sharing
enhancement
Facebook Debugger
142
Page titles unique, under 60 characters, and include target keyword naturally
important
Screaming Frog
143
Meta descriptions unique and under 155 characters
enhancement
Screaming Frog
144
H1 tag present, unique per page, and matches the page's primary topic
important
Screaming Frog
145
Images have descriptive alt text — not keyword-stuffed, not empty on content images
important
Screaming Frog alt audit
146
CMS-generated tag and category pages reviewed for thin content — improved or noindexed
important
Manual review
147
Quarterly audit calendar established — this document is run again every 90 days
critical
Project management
50%
of websites have critical technical SEO issues actively hindering search performance — yet most never fix the highest-leverage items first.
— 2026 industry survey, aggregated
The #1 crawlability mistake
Infinite crawl traps from faceted nav
An e-commerce site with 1,000 products and 5 filterable attributes can inadvertently generate over a million low-value URLs. Googlebot wastes its crawl budget on the combinatorial explosion, and the real product pages get crawled less than they should. The fix is robots.txt disallow rules for parameter combinations that don't produce unique content, plus canonicals pointing to the unfiltered category.
The schema mistake that triggers manual actions
FAQ & Review schema misuse
Adding FAQ schema to pages that don't visually have an FAQ section — or adding Review schema with manufactured ratings — is a policy violation Google increasingly enforces with manual actions. Implement schema only when on-page content supports it, exactly as visible to users.
03 / Core Web Vitals · 2026 thresholds
Four metrics, three zones. Where your site lands at p75 mobile.
All measured at the 75th percentile of mobile field data from CrUX, never lab tests. Missing any one threshold puts you in "needs improvement" territory; competitive queries now require passing all three.
LCP
2.5s
Largest Contentful Paint
Time until the largest visible element (hero image, H1) is rendered.
Good≤ 2.5s
Needs work2.5–4.0s
Poor> 4.0s
INP
200ms
Interaction to Next Paint
Responsiveness across all interactions during a session — not just the first.
Good≤ 200ms
Needs work200–500ms
Poor> 500ms
CLS
0.1
Cumulative Layout Shift
Visual stability — how much content shifts unexpectedly during load.
Good≤ 0.1
Needs work0.1–0.25
Poor> 0.25
TTFB
800ms
Time to First Byte
Server response time. Foundational — every other metric inherits from it.
Good≤ 800ms
Needs work800ms–1.8s
Poor> 1.8s
”
One LA client — an e-commerce site on WordPress with heavy WooCommerce customization — came to us with 1,400ms INP at p75. We refactored the product filter with debouncing and scheduler.yield(), moved non-critical third-party scripts to load after user interaction, and reduced category-page JavaScript by 40%. INP dropped to 180ms within two weeks. Mobile conversion rose 18%.
— Olibro engagement note, 2025
04 / Run it on your own site
Two days. Twelve hours. The exact schedule we follow internally.
147 checks looks daunting. Most are fast. Day one is data collection — you can run the long crawl in the background. Day two is scoring, prioritizing, and presenting the top ten fixes.
Day 1
Data collection
6 hours
Morning2 hours
Full Screaming Frog crawl
Configured to render JavaScript, respect robots.txt, capture all technical data. Export to CSV. On a 5,000–50,000-URL site, expect 30–90 minutes of crawl time you can run in the background.
Midday2 hours
Pull GSC data
Core Web Vitals, Page Indexing, Crawl Stats, Enhancements, International Targeting, last 90 days of Performance. Export everything as CSV/Sheets.
Afternoon2 hours
PageSpeed + Rich Results + AI fetches
PSI on the top 10 page types. Rich Results Test on one example of each template. Fetch robots.txt, llms.txt, sitemap.xml. Test 5 URLs as Googlebot Smartphone in URL Inspection. Test 3 URLs as GPTBot via curl.
Day 2
Scoring & prioritization
6 hours
Morning3 hours
Walk all 147 checks
For each, mark Pass / Fail / N/A based on Day 1 data. For Fails, note the observed value and the gap from target (e.g., "INP is 340ms at p75, target 200ms, gap 140ms").
Afternoon3 hours
Sort by ROI, build the fix list
Sort Fails by tier (Critical → Important → Enhancement). Estimate hours for each. Group by owner (dev, content, SEO, design). Present the top 10 fixes to the client or internal team.
Our secret weapon
We maintain a living template of this audit as a Google Sheet with dropdown fields for Pass/Fail, auto-calculated priority scores, and automatic owner assignment. It takes the cognitive load out of running 147 checks and lets us focus on the interesting findings, not the mechanical scoring.
05 / What order to fix things in
An audit produces 40–80 issues. You can't fix all of them this quarter.
Here's the sequence that delivers the biggest ranking lift in the shortest time, based on what we've observed across 270+ client engagements. Click any phase to expand.
01
Days 1–7
Visibility emergencies
+
Days 1–7
Visibility emergencies
02
Days 8–30
Indexation & rendering
+
Days 8–30
Indexation & rendering
Clean up the index. Fix "Discovered – currently not indexed", remove thin content, consolidate duplicates via canonicals, add self-referencing canonicals everywhere. Verify critical content renders in initial HTML. Test with Googlebot Smartphone and GPTBot UA. Typically recovers 10–30% of lost rankings on a neglected site.
03
Days 31–60
Core Web Vitals
+
Days 31–60
Core Web Vitals
Address LCP, INP, and CLS in that order. LCP is usually the easiest to fix (image optimization, font preloading, above-the-fold lazy loading). INP is the hardest and most valuable (JavaScript refactoring, long task breaking, third-party script auditing). CLS is fast (explicit dimensions, reserved ad space).
04
Days 61–90
Structured data & schema
+
Days 61–90
Structured data & schema
Add Organization, Article, FAQ, LocalBusiness, Product, and Breadcrumb schema across all relevant page types. Validate everything in Rich Results Test. Rarely delivers immediate ranking wins but meaningfully improves AI citation rates and SERP real estate over 3–6 months.
05
Ongoing
Enhancement & refresh
+
Ongoing
Enhancement & refresh
Remaining Enhancement-tier items get scheduled across quarterly refreshes. llms.txt refinement, international SEO edge cases, security header hardening, unused JavaScript tree-shaking, HTTP/3 enablement. None moves rankings by itself — collectively they're what separates the top 5% from merely adequate.
Sources & methodology
Standards and data points cited in this audit are drawn from these primary sources, accessed January–April 2026.
- Google Search Central documentation, 2024–2026 — Core Web Vitals, mobile-first indexing, JavaScript rendering, structured data guidelines.
- Google's December 2025 Rendering Update announcement on non-200 status code exclusion from the rendering pipeline.
- web.dev research on Core Web Vitals, including the 24% abandonment reduction figure for pages passing all three metrics.
- DebugBear and PageSpeed Insights documentation on field vs. lab data interpretation.
- Schema.org vocabulary and Google Rich Results Test guidelines for JSON-LD implementation.
- llmstxt.org specification for AI crawler content hints.
- OpenAI, Anthropic, and Perplexity platform documentation on respective crawler user-agents and content access policies.
- Patchstack Annual Threat Report for WordPress vulnerability data.
- Qualys SSL Labs grading methodology for TLS configuration assessment.
- Olibro Design internal engagement data, 2010–2026, aggregated and de-identified — 270+ client sites across healthcare, e-commerce, nonprofit, and professional services.
Jelena Misić
SEO Specialist
I help businesses get found online—whether they're selling products, services, or serving local communities.
More posts by Jelena Misić
Have a project in mind?
We help businesses ship websites that work — on WordPress, Shopify, and AI-ready infrastructure.