Synthesis
Tier-by-Tier Comparison
What each audit method caught, what it missed, and where all three converge. Convergence signals highest-priority actions.
Audit A
Koray Semantic SEO
Topical authority model, entity coverage, semantic gaps, content architecture. Pure methodology, no external data integration.
Audit B
Koray + Merlino Stack
Koray base augmented with Bill Slawski patent research, Matt Diggity on-page signals, local SEO overlays, and DataForSEO competitive data. Three parallel lanes.
Audit C
Full Ecosystem
All lanes: technical, semantic, local, links, structured data, competitor topology, growth opportunities. Broadest coverage.
Audit Compounding Value: What Each Tier Added
Client: TurnKey Bath Remodel Date: 2026-06-11 Purpose: Show the exact delta in diagnostic precision and actionability at each audit tier.
How to Read This Document
Each row is a finding. The columns show which tier first surfaced it and what deeper tiers added. "Caught by A only" means a single-methodology approach would have found it. "Required B to surface" means pure Koray missed it. "Required C to resolve" means neither A nor B had the forensic capability to tell the client what to actually do.
The pattern you will see: Tier A finds structural problems. Tier B finds commercial priority and root cause. Tier C tells you exactly what happened, who did it, and what the fix costs.
Section 1: What Audit A (Pure Koray Semantic SEO) Caught
Audit A score: 24/100. Methodology: Koray Tugberk's entity-attribute-value topical authority framework, semantic content networks, and site-wide ngram consistency doctrine.
| Finding | What A Found | A's Confidence |
|---|---|---|
| Entity ngram inconsistency | "bath remodel," "bathroom remodel," "bathroom renovation" used interchangeably sitewide | HIGH: direct textual observation |
| 13 missing core-section entity-attribute pairs | No cost data, no acrylic attribute, no one-day install process node, no Medicare/insurance attribute on walk-in tubs | HIGH |
| City pages are copy-paste templates | Only city name swapped, Slidell lists fabricated services TurnKey does not offer | HIGH: direct page read |
| Showers page H1 duplicated as H2 | Broken contextual vector | HIGH |
| Meta descriptions absent on homepage, /showers/, /bathtubs/ | Three critical IR zones missing | HIGH |
| Sister domain likely same company | Splitting topical authority | MEDIUM: suspected, not confirmed |
| New Orleans city page 404 | Most important geo-entity absent | HIGH |
| Title-to-H1 macro context misaligned on hub pages | "Shower Remodeling in New Orleans" vs. "Shower Services" | HIGH |
| Internal linking anchor text violations | 14 city pages with identical anchors to same destinations | HIGH |
| Publishing momentum below threshold | ~2 posts/month, outer-section topics before core-section nodes | HIGH |
Audit A blind spots: A could not confirm whether the sister domain was the same legal entity, could not diagnose why the GMB was invisible, could not explain the 94% broken backlink figure (which turned out to be a WAF artifact), could not quantify the link profile risk, could not evaluate the AI crawling failure, and could not estimate conversion impact of the DIY blog post traffic mismatch.
Section 2: What Audit B Added That A Missed
Audit B added three lanes: Semantic + Merlino Stack (score 29/100), Local/GBP (score 21/100), and Patents (score 18/100). Methodology additions: Slawski Google patent database, Diggity on-site rules, Merlino local SEO stack, GBP/DataForSEO API, intent-behind-intent analysis.
What B Added on Top of A
| Finding | A's View | B's Addition |
|---|---|---|
| Sister domain | Suspected same company, split authority | CONFIRMED same GoHighLevel infrastructure; DataForSEO found TWO active GBP listings (CIDs confirmed); sister ranks positions 8+11 for money keyword with NOLA-keyword-optimized title; 301 redirect directive issued |
| GMB invisibility | Not in scope for A | ROOT CAUSE IDENTIFIED: GMB pinned at 1401 Distributors Row 70123 (industrial park) while website/schema shows 3436 Magazine St 70115, different streets, different ZIPs, different neighborhoods; entity conflict breaks prominence scoring |
| Four-way NAP conflict | Not in scope for A | QUANTIFIED: two phone numbers, two addresses, press releases/Facebook/bestremodel.com on GMB NAP, website/schema on different NAP; Google cannot resolve single entity |
| Homepage H1 keyword absence | A flagged H1 as "emotional marketing tagline" | B applied Diggity three-kings rule specifically: title, H1, and first paragraph must carry primary keyword; gave exact rewrite formula ("New Orleans Bathroom Remodeling Completed in 1 Day") |
| Walk-in shower page intent mismatch | A flagged attribute hierarchy violation | B identified the hidden intent group missed entirely: ADA/aging-in-place, mobility anxiety, care identity driver; recommended specific H2 targeting that dominant hidden intent |
| City page civic H2s | A flagged as template copy | B identified these as commercial topical coherence violations per Koray contextual domain mismatch rule; specified the replacement content (permit requirements by parish, local remodeling stats, post-Katrina housing stock characteristics) |
| Tub-liner page | A did not specifically address | B identified as highest-traffic asset and flagged cost data ($800-2,500) buried mid-page; gave snippet formula for opening paragraph; added New Orleans humidity as entity-attribute differentiator |
| AI/LLM visibility | Not in scope for A | B ran ai-llm-seo checklist: all 5 pages fail "key answer in first 150 words" and "FAQPage JSON-LD present"; cited per-engine citation preference table |
| FAQPage schema gap | A did not evaluate schema | B confirmed zero JSON-LD schema on 4 of 5 fetched pages; gave exact field list for FAQPage deployment (cost, Medicare coverage, timeline, aging-in-place) |
| Backlink profile risk | A did not evaluate links | B applied temporal profile patent (US 7,346,839): press-release syndication footprint matches PBN cluster-deletion signature; 440 organization + 103 news platform links flagged; US 8,117,209B1 Reasonable Surfer Model analysis showed near-zero equity transfer |
| Publishing momentum | A flagged volume issue | B added specific content titles for a 6-article launch batch and applied Koray Momentum doctrine (patternless publish over 7-10 days) |
B's residual blind spots: B could not confirm whether the 691 broken links were real or a tool artifact. B could not determine if sister domain ownership was the same legal entity or just the same infrastructure. B could not evaluate the homepage JS-rendering issue (no technical crawl capability). B could not assess CRO conversion elements or competitive gap specifically.
B increased the diagnostic value from "here are structural problems" to "here is why they are structural problems and what the client should specifically do about each one."
Section 3: What Audit C Added That B Missed
Audit C added seven specialized lanes. Each added forensic or specialist precision that changed the recommended action.
C.1 Forensics (Score 71/100) - Highest Confirming Value
| Finding | What B Had | What Forensics C Added | Action Delta |
|---|---|---|---|
| 691 broken backlinks | Treat as real broken links, disavow recommendation | CONFIRMED: WAF artifact. Every "broken" link returns HTTP 200 to browsers. DataForSEO's crawler hit WPX Cloud's bot-blocker. Domain has a data-quality problem in all backlink tools, not an actual link crisis. | REVERSES the disavow-691-links recommendation. Client does NOT have a broken link crisis. Saves weeks of panic remediation. |
| Sister domain ownership | Suspected same company | CONFIRMED same operator beyond reasonable doubt: identical WPX Cloud /24 subnet (194.1.147.x), PHP 8.3.31, GoHighLevel (LeadConnectorHQ) forms, identical URL structure, identical 14 city service area list. | Changes 301 redirect from "recommend" to "execute immediately, confirmed same operator." |
| Press release footprint | Patent-identified risk | CONFIRMED: all 440 organization + 103 news links trace to a single Press Advantage press release campaign ("50-Year Warranty" announcement). Three target URLs only (homepage, /contact/, /bathtubs/bathtubs-replacement/). Zero links to any city page. | Confirms press release campaign was the entire link strategy and it went to the wrong pages even when live. |
| Live backlinks | Not evaluated per-page | CONFIRMED: all 36 live backlinks concentrate on 3 pages; zero external links to any of 14 city pages, /showers/, /faq/, or /our-projects/. | The pages intended to drive local rankings have no link equity whatsoever. Redirects context for link-building to city pages specifically. |
| /gallery/ 404 | Not in scope for B | CONFIRMED via Wayback CDX: returned 200 on 2024-10-14, now gone with no 301 redirect. | Adds a specific 301 redirect action that B could not surface. |
| GBP UTM URL | B confirmed CID exists | Confirmed GBP was crawled by DataForSEO on 2025-12-20 via UTM URL, proving the listing is active but not prominent. | Confirms the listing is not suspended; the visibility gap is prominence-based, not suspension-based. |
C.2 Technical (Score 22/100)
| Finding | What B Had | What Technical C Added |
|---|---|---|
| Homepage JS rendering | Not detected | SSR score 0/100. curl confirmed: 127KB HTML but only 1,471 visible characters post-script-strip. AI crawlers see empty page. WP Rocket IE-compatibility block is the first 150+ words of HTML. |
| OG image is favicon | Not detected | og:image = favicon.png (16x16). Twitter card is "summary" not "summary_large_image." Brand name inconsistency in OG tags. |
| /projects/ 404 | Not detected | Returns 404. Correct URL is /our-projects/ (HTTP 200, in sitemap). /projects/ not in sitemap but may receive referral traffic. |
| Security surface | Not in scope | wp-login.php returns HTTP 200. No security headers. xmlrpc.php URL advertised in HTML head. |
| Server-side positives | Not quantified | TTFB 0.159s (excellent). HTTP/3 + QUIC enabled. These are competitive advantages to preserve. |
C.3 Schema (Score 40/100)
| Finding | What B Had | What Schema C Added |
|---|---|---|
| @id not GMB CID | B flagged generically | CONFIRMED across all 7 pages fetched. Specific field names: @id, mainEntityOfPage, additionalType, knowsAbout, areaServed Wikidata Q-numbers all absent. |
| hasOfferCatalog gap | Not in scope for B | Only one generic Offer listed despite 6 distinct services. Google cannot associate the entity with specific service types. |
| Metairie phantom entity | Not detected | City page spawns a second LocalBusiness @id node using the New Orleans address (not Metairie address). Creates a phantom duplicate entity in the graph. |
| Duplicate WebSite block | Not detected | Homepage has one WebSite block with no @id, creating parser ambiguity. |
| Em dash in warrantyScope field | Not detected | Two em dashes in schema field. Violates no-em-dash rule. Must be stripped. |
| FAQPage location specifics | B said FAQPage needed on service pages | C confirmed exactly which pages (showers, walk-in-tubs) and which questions (cost, Medicare, timeline, aging-in-place) per the 5-question minimum. |
C.4 Backlinks (Score 12/100)
| Finding | What B Had | What Backlinks C Added |
|---|---|---|
| Press release footprint | Patent risk identified | QUANTIFIED: 138 xpr-gannett.com links, 39 marketminute.com links, 30+ identical mybluehost.me subdomain entries. Anchor repeats 138 times. Spam anchors including t.me/s/quarterlinks25 and BuySeoLink.com injected. |
| Referring domain gap | General deficit noted | Specific numbers: 188 RDs (TurnKey) vs. 330 (ezbaths.com) vs. 2,711 (rebath.com) vs. 4,523 (bathfitter.com). At 5 new RDs/month, closing the ezbaths.com gap takes 28 months. |
| Dofollow ratio anomaly | Not detected | 96% dofollow vs. healthy 60-80% benchmark. Absence of naturally nofollowed links is itself a footprint signal. |
| Domain rank paradox | Not quantified | DR 233 despite 699 backlinks vs. ezbaths.com DR 211 with only 504 backlinks. Quality signal suppressed by press release footprint. |
C.5 Competitors (Score 49/100)
| Finding | What B Had | What Competitors C Added |
|---|---|---|
| EZBaths as primary competitor | Known | Specific attack vector: EZBaths ranks #17-22 for "bathroom remodeling new orleans" cluster with generic city page. TurnKey has a physical New Orleans address advantage. A hyper-local NOLA page displaces EZBaths specifically. |
| bigeasybathrooms.com threat | Not identified | Shares 7 of TurnKey's 14 service cities with ETV of $6 and thin templated city pages. Immediately beatable with substantive content in any shared city. Slidell already in TurnKey's top-5 traffic pages = start there. |
| Cost content gap | Not quantified | "How much to replace a tub" (vol 1,000) and "fiberglass tub replacement cost" (vol 1,000) both held by bigeasybathrooms.com with thin content. TurnKey not ranked. Single cost-guide page captures this immediately. |
| EZBaths URL architecture | Not evaluated | /city/service/ nested URL pattern is the proven local ranking architecture in Louisiana. EZBaths' /baton-rouge/walk-in-tubs/ ranks #4-6. TurnKey has no city+service nested pages. |
| Seniors/aging-in-place gap | B identified hidden intent | C quantified: EZBaths ranks for "bathtubs for seniors" and "elderly walk-in tub" at vol 5,400 each. TurnKey has zero accessibility content. Highest-volume adjacent cluster with no TurnKey presence. |
| Financing gap | Not detected | TurnKey has no financing page. EZBaths (18-month no-interest) and all nationals rank for financing terms. |
C.6 Content Quality (Score 44/100)
| Finding | What B Had | What Content C Added |
|---|---|---|
| Zero named authors | B noted E-E-A-T gap | C confirmed zero named authors sitewide. No team page accessible from navigation. All 9 testimonials use generic names without photos or linked Google profiles. Specific AI citation consequence: ChatGPT, Perplexity, and AI Overviews prefer named expert sources. |
| DIY blog issue | B flagged intent mismatch | C confirmed it is 52% of all ranked keywords (25 of 48). Added specific fix: author schema with contractor credentials plus conversion bridge section. |
| External citations absent | Not detected | Zero external citations across all 6 audited pages. No links to manufacturer specs, NARI standards, LSLBC license verification. |
| No publish dates | Not detected | No publish dates or last-updated dates on any content. Missed freshness signal and featured-snippet opportunity. |
| City page tourist trivia specifics | B flagged generically | C provided specifics: Lafreniere Park, Lakeside Shopping references confirmed in Metairie page. Named the specific replacement content needed (local permit requirements, local remodeling stats, popular styles for post-Katrina housing stock). |
C.7 AI Visibility (Score 38/100)
| Finding | What B Had | What AI Visibility C Added |
|---|---|---|
| llms.txt staging contamination | Not detected | All internal URLs in llms.txt point to vm34.wprise.ai. Five-minute fix with massive AI crawling impact. |
| WP Rocket blocking AI content extraction | Not detected | WP Rocket IE-compatibility JavaScript block is the first 150+ words of HTML on all pages. AI crawlers see code, not content. has_intro_summary scores 0/1 on every page. |
| FeatGEO dimension scoring | Not quantified | Trust dimension caps at 8.3/25 sitewide. quotation_level = 0 and cite_sources_level = 1 identified as the two primary gaps. |
| Heading structure server-rendering | Not detected | Showers page and FAQ page have zero H2/H3 in server-rendered HTML. Tub liners page scores 25.0/25 on Structural (full H2+H3 hierarchy) as the positive benchmark. |
C.8 CRO (Score 54/100)
| Finding | What B Had | What CRO C Added |
|---|---|---|
| Modal-only form | Not in scope | Confirmed modal-only across all pages, no inline fallback. Modal-triggered CTAs fail on slow mobile. |
| Three competing CTAs | Not detected | Three CTAs above fold confirmed: Design Consultation, Call, FREE Estimates. Decision paralysis. |
| Social proof placement | Not detected | Testimonials below the fold on all inner pages. Homepage carries 4.9-star trust bar above fold but inner commercial pages do not. |
| 1-day install USP underused | B flagged entity consistency issue | C confirmed 1-day USP buried mid-page on /showers/ and absent from all city page H1s. Formula recommended: "[Service] in [City] - Done in 1 Day." |
| /contact/ headline failure | Not detected | "Get In Touch" fails the five-second clarity test. Specific replacement given. |
| Sticky bar gap | Not detected | Sticky bottom bar confirmed active and correct. Gap: no "Text Us" SMS option. 35% of mobile users prefer SMS for initial contact. |
Summary: Compounding Value Table
| Dimension | Findings A Alone | Findings B Added | Findings C Added | Net Findings |
|---|---|---|---|---|
| On-page / entity | 10 | 6 (deeper root cause) | 3 (JS rendering, OG, security) | 19 |
| Local / GMB | 0 | 9 (root cause + API confirmation) | 2 (photo count, hours conflict) | 11 |
| Links / authority | 1 (suspected PR links) | 4 (patent risk, broken % concern) | 6 (WAF revelation, per-page gap, spam anchors) | 11 |
| Schema | 0 | 4 (field list) | 6 (phantom entity, duplicates, em dash) | 10 |
| Technical | 0 | 0 | 7 (SSR, llms.txt, gallery 404, security) | 7 |
| Competitive | 0 | 2 (sister domain context) | 6 (specific competitor gaps, city architecture) | 8 |
| Content quality | 1 (publishing cadence) | 3 (intent analysis, snippet positioning) | 6 (authors, citations, dates, specifics) | 10 |
| AI / LLM | 0 | 3 (checklist, FAQPage) | 5 (WP Rocket, llms.txt, FeatGEO scores) | 8 |
| CRO | 0 | 0 | 8 (form, CTAs, proof placement, SMS) | 8 |
| Total | 12 | +31 | +51 | 94 |
The compounding argument: Tier A alone gives a 12-finding diagnosis. You can tell the client they have semantic problems. You cannot tell them why their GMB is invisible, whether their broken links are real, what their AI crawling failure looks like, or what their competitors are specifically doing that is working. Tier B brings the diagnosis to 43 findings with commercial prioritization and root-cause specificity. Tier C brings it to 94 findings and, critically, REVERSES a major recommendation (the broken link disavow), confirms the sister domain as the same operator (moving it from "probable" to "actionable"), and adds the seven-lane specialist perspective that turns findings into a work order.
No fabricated data. All evidence cited from Phase 0 DataForSEO pull, confirmed via per-lane live fetches, Wayback CDX, and DataForSEO API calls per method notes in each lane report. 2026-06-11.