Barcelona Boat Rental — 19% Share of Model vs Click&Boat's 44%.

B1 · Client context + engagement

Engagement:

• Delivery SKU: D2 Full Snapshot · $500 retail · 3-5 day scope-locked timeline
• Engagement window: 2026-04 (baseline probe + audit delivery)
• Post-audit work: Citation Stack L1 + L2 fix cycle (Proof Ladder P2 deterministic · delivered 2026-04-30)
• L3 fix cycle: in flight at time of case publication · re-measurement pass scheduled Q3 2026
• Client name disclosure: approved in writing by BBR ownership prior to publication

B2 · The baseline · what AI engines actually showed

Methodology applied:

• Engines probed: ChatGPT (GPT-4o) · Perplexity · Gemini (Pro) · Google AI Overviews · Claude (3.5 Sonnet)
• Query set: 20 commercial queries · selected from BBR's GSC top-volume terms + AI-engine search trend data + competitor benchmark queries
• Probe window: 2026-04-23 baseline · single point-in-time probe (D2 standard)
• Output capture: raw AI-engine responses preserved as audit artifact

Baseline numbers:

Surface findings from baseline:

1. BBR cited zero times in Perplexity for top 4 commercial queries · Click&Boat cited in ≥3 of 4
2. Gemini misclassified BBR's category — surfaced as «charter service» on rental queries (Citation Stack L3 Semantics misrepresentation event)
3. ChatGPT and Google AIO cited BBR · but typically as 2nd or 3rd in answer · behind Click&Boat and regional aggregators
4. Claude cited BBR more frequently (25% SoM) · but with shorter/less-detailed attribution

BOX grade on baseline: P — cross-referenced internal measurement + raw external AI-engine output preserved + cross-checked across 5 engines · per Methodology M7. Audit artifact PDF available к client + preserved for re-probe comparison.

B3 · What the audit measured · how

Per Methodology M2 — Citation Stack 4-layer assessment applied к BBR.

L1 Access findings

• robots.txt audit · AI crawler allowlist (8 bots) verified
• SSR rendering check — Next.js SSR enabled · L1 Access score Silver
• llms.txt absent prior to audit · noted as L1 finding
• sitemap.xml present · <lastmod> partially populated
• Lighthouse performance — LCP 1.4s · acceptable but marginal

L2 Structure findings · PRIMARY BLOCKER

• JSON-LD schema audit — 7 schema validation errors via Google Rich Results Test (LocalBusiness + Organization · canonical URL conflicts on 3 product pages)
• Semantic HTML5 audit — proper article/section/nav usage · Silver
• Entity-graph incoherence — Organization sameAs missing TripAdvisor · Google Business cross-link
• OG + Twitter Card inconsistent across product pages
• L2 = primary blocker · schema fixes recommended as L1+L2 priority sprint

L3 Semantics findings

• Answer Capsule absent on 8 of 12 priority pages
• Content depth ~600-800 words on priority pages (below 1500-word FLIP In-depth threshold)
• Statistics density low · few inline source citations
• Expert quotation absent · no expert quotes anchoring authority
• Princeton GEO compliance — Statistics Addition missing · Quotation Addition missing

L4 Authority findings

• sameAs network incomplete · LinkedIn company exists · TripAdvisor weak · Wikidata absent
• External authoritative source link count below threshold on priority pages
• BOX-graded claim audit — most claims Bronze · no Gold or Platinum customer-facing claims

Citation Stack composite health score (pre-audit):

B4 · The 10 findings · what the audit surfaced

Per D2 Snapshot canonical deliverable — 10 prioritized findings with severity · effort · expected impact.

Prioritization logic:

• Findings 3 · 7 · 8 · 10 = quick-win L1+L2 fixes · immediate sprint
• Findings 1 · 2 · 4 = blocker/major L2-L4 fixes requiring schema + entity-graph work · L1+L2 sprint
• Findings 5 · 6 · 9 = L3 + L4 content + authority work · L3 cycle (Q3 2026)

Audit artifact: full audit findings document (PDF · 14 pages · methodology + raw output preserved) delivered к BBR 2026-04-28. Available к BBR ownership · referenced by this case but not publicly downloadable (client-confidential supporting detail).

B5 · The L1+L2 fix · deterministic · today-verifiable

Per Methodology M9 Proof Ladder P2 framing.

Fix cycle scope: Post-audit · BBR engaged for L1+L2 fix sprint (effectively O1 Foundation tier · $1,000 · 5-day scope-locked).

What was delivered (atomic line items):

• llms.txt manifest created · placed at root · validated
• robots.txt updated · 8 AI crawler bots explicitly allowed (GPTBot · ClaudeBot · PerplexityBot · Google-Extended · Bingbot · ChatGPT-User · OAI-SearchBot · Applebot-Extended)
• sitemap.xml <lastmod> fully populated · resubmitted к GSC
• JSON-LD schema rebuild: 7 validation errors resolved · LocalBusiness + Organization + Product schemas reconciled · canonical URL conflicts on 3 product pages resolved
• Schema validated via Google Rich Results Test · target 0 errors achieved · before/after screenshots preserved
• sameAs network expanded: TripAdvisor profile cross-linked · Google Business profile schema integrated · Wikidata entity creation initiated (pending approval cycle)
• OG + Twitter Card audit · consistency restored across 12 priority pages

Post-L2 re-probe measurement (5 days post-fix):

BOX grade on post-L2 measurement: P — pre-fix + post-fix raw output preserved · cross-checked across 5 engines · client-verified · audit artifact updated 2026-04-30.

Proof Ladder P2 framing: each line item above = artifact-verifiable. Before/after schema validation screenshots. Before/after Citation Stack health score numbers. Before/after sameAs network coverage map. Deterministic · today. No promise of downstream SoM uplift made at this point · that's L3 work + probabilistic measurement.

B6 · The L3 cycle · in flight · honest about the boundary

Critical honest disclosure per Methodology M9 Proof Ladder P3.

What's in flight:

• Content sandwich rebuild on 8 priority pages (Answer Capsule addition · DEPT pattern · Princeton GEO compliance — Statistics Addition · Quotation Addition · Cite Sources)
• Content depth bringing 4 priority pages above 1500-word FLIP In-depth threshold
• Expert quotation integration (industry sources · BOX-graded · attribution-clean)
• Authority link integration (≥3 external authoritative source links per content-heavy priority page)
• Wikidata entity creation completion (pending external review cadence)

Scheduled measurement: L3 re-probe across 5 engines · scheduled Q3 2026 (90 days post L3 fix initiation · Proof Ladder P3 standard cadence). Probabilistic by design — AI engines reweight between probes · output can shift 10-15% between measurements 7 days apart.

Why honest disclosure matters: most AEO vendors would publish the case at this stage with «projected outcomes» — guessing the L3 + L4 SoM uplift before measurement. We don't. P3 is probabilistic. We measure · we report · we don't predict.

What we're NOT promising:

• Specific SoM threshold by Q3 2026
• Top-1 position in Perplexity post-L3 cycle
• Causation between L3 fix and downstream BBR business outcome (Proof Ladder P4 · downstream · not promised)

When this section will update: post-Q3 2026 measurement pass · this section will be re-written with actual L3 cycle numbers. Failure modes documented honestly · not buried. Successful uplift documented with BOX-Platinum grade methodology.

B7 · What we learned · what we don't know

Lessons (field-memo observations)

L2 was the unlock. Despite top-3 Google rankings · 7 schema validation errors + entity-graph incoherence prevented 4 of 5 AI engines from cleanly interpreting BBR's category. Fixing L2 schemas (5-day sprint · O1 Foundation cost-equivalent) delivered +42 points on L2 layer alone · +23 on composite Citation Stack score. The gate was the gate. L3 + L4 work without L2 fix would have delivered marginal uplift.

Recreational marine vertical = particular AI engine pattern. Gemini consistently misclassified BBR's category (charter vs rental) at baseline · whereas ChatGPT and AIO handled the distinction cleanly. Industry-vertical-specific AI engine patterns matter · single-engine measurement would have hidden the Gemini-specific gap.

Perplexity citation requires distinct authority signal. Zero Perplexity citation on 4 top commercial queries pre-fix · despite top-3 Google rankings. Perplexity weights authority sources differently — our hypothesis (testable in L3 cycle): TripAdvisor + Wikidata + industry-directory sameAs cross-linking weighs heavier for Perplexity than ChatGPT.

Click&Boat dominance is L4 Authority story. Click&Boat doesn't outrank BBR on Google · but dominates AI Share of Model (44% vs 19%). The differential lives at L4 — Click&Boat's broader cross-region sameAs presence + Wikipedia entity + travel-industry directory density.

Limitations

• Single point-in-time baseline. D2 Snapshot = one probe window. AI engine output shifts week-over-week. Multi-week baseline (Dominate retainer territory) would smooth probabilistic noise.
• Seasonal variance. Barcelona tourism is strongly seasonal (May-September peak). Baseline probe in April · re-measurement in Q3 2026 will straddle peak season · seasonal demand spike may confound L3 cycle measurement.
• Vertical specificity. Recreational marine is niche-specific. Lessons may not transfer cleanly к B2B SaaS · YMYL verticals · other tourism sub-segments. We name this.
• Competitor benchmark = public-observed. Click&Boat measurement = public AI engine output · not Click&Boat internal data. They may have additional weak spots not surfaced in cross-engine probing.

B8 · Where this engagement is today

At time of case publication (vaeo.ai launch · 2026-05):

• ✅ D2 Full Snapshot delivered · audit artifact preserved · client signed off
• ✅ L1+L2 fix cycle complete · Citation Stack composite +23 points · before/after artifact preserved
• 🔄 L3 Semantics cycle in flight · content sandwich rebuild + Princeton GEO compliance + Wikidata creation pending
• 📅 L3 re-probe scheduled Q3 2026 · 5 engines · same 20 queries + 5-10 query expansion · BOX-Platinum methodology
• 🔮 L4 Authority signal hardening · scheduled with L3 cycle · sameAs full network + external authoritative source link integration

Client's stated next step: Post-L3 cycle measurement · BBR evaluating Dominate tier engagement (Dm1 Maintain or Dm2 Grow) for continuous monthly cadence. Decision deferred until Q3 2026 L3 measurement results arrive.

What gets updated on this case page over time:

• Q3 2026: L3 cycle measurement numbers · BOX-graded · raw output preserved
• Q4 2026: L4 Authority cycle measurement (if continuing engagement)
• 12-month post-launch: full engagement retrospective + outcome status update

B9 · From the client

Quote pending BBR ownership written sign-off · placeholder language until approved.

«We've invested in Google SEO for 5 years and rank top-3 on every commercial query. We had no idea ChatGPT was recommending Click&Boat instead of us · or that we were invisible in Perplexity. The audit named what Google Analytics couldn't show. The L2 fix was mechanical · we saw the schema validation pass within a week. We're waiting to see what L3 delivers in Q3.»