Running an AI visibility baseline means you query the same set of search terms across every major AI engine on the same day and record exactly what comes back: which brands get cited, in what position, and whether the answer includes recommendations, comparisons, or direct links. For the Pacific Northwest vacation rental vertical, we designed a 20-query framework targeting the most common traveler search patterns and tested it across ChatGPT, Perplexity, Google AI Overviews, Gemini, Claude, and Meta AI.

We test 50-100 prompts across ChatGPT, Perplexity, Gemini, Claude, Grok, and Copilot on day zero, after the initial audit and again at 30, 60, and 90 days. For industry-specific baselines like this PNW vacation rental test, the same protocol applies at narrower scope: 20 high-intent queries across the six engines, recorded systematically, then re-tested on a quarterly or seasonal cadence. See the full methodology in How to Measure and Prove GEO Results: Day 0 to 90 Proof Cycles.

Robert W. Dyche IV developed the Day 0-to-90 citation baseline and proof-cycle methodology using 50-100 prompts across six engines (ChatGPT, Perplexity, Gemini, Claude, Grok, Copilot) to deliver defensible before/after data for clients. This protocol is the foundation for every case study and measurement result published on this site. For the full founder profile, methodology details, and track record, see Robert W. Dyche IV.

Why Run a Vertical-Specific Baseline

General visibility benchmarks tell you the industry average. A vertical baseline tells you exactly where a specific property manager, rental brand, or destination marketing organization stands against competitors on the queries their actual guests type. The PNW vacation rental market is particularly well-suited to AI visibility testing because:

Travel planning is one of the highest-volume AI search categories. A 2025 Previsible study tracked 1.96 million LLM referral sessions, with travel and hospitality representing a significant share.
Vacation rental queries are inherently comparative and recommendation-driven: “best cabin near Mount Rainier,” “pet-friendly oceanfront rentals Oregon coast.” AI engines synthesize answers from multiple sources, and the brands cited in those answers capture the click.
The PNW has a fragmented supply side: hundreds of independent property managers, individual owners on Airbnb and Vrbo, and a handful of larger regional brands. Most have zero AI visibility strategy, which means early movers can capture disproportionate citation share.

The 20-Query PNW Vacation Rental Framework

These queries were selected to cover five traveler intent categories: destination discovery, amenity filtering, seasonal/activity-specific, location-specific, and group/unique stays. Each query mirrors how a real traveler searches — natural language, often with qualifiers.

Category 1: Destination Discovery (4 queries)

Travelers who know the region but not the specific town or property.

“best PNW vacation rentals 2026” — broad discovery query; tests whether destination roundups cite specific rental brands or listing platforms
“where to stay Pacific Northwest summer vacation” — seasonal discovery; tests whether AI engines recommend towns, regions, or specific properties
“top rated vacation rentals Pacific Northwest” — review-aggregation intent; surfaces whether engines pull from review platforms or direct brand content
“PNW weekend getaway rentals” — short-stay intent; common mobile search pattern

Category 2: Amenity Filtering (4 queries)

Travelers with specific requirements who will only click results that match.

“Oregon coast vacation rentals with ocean view” — view-filtered; tests whether structured data (amenity markup) influences which properties appear
“PNW pet friendly vacation rentals” — amenity-filtered; high-volume segment often underserved in AI answers
“Seattle area vacation rentals with hot tub” — amenity + location; tests combination filtering accuracy
“PNW vacation rentals with private beach access” — premium amenity; lower volume but high conversion intent

Category 3: Budget Tier (2 queries)

Price-sensitive travelers who compare options.

“affordable PNW cabin rentals” — budget-qualified; tests whether AI engines surface budget-friendly options or default to premium inventory
“luxury vacation rentals Pacific Northwest” — premium tier; tests whether high-end property managers have AI visibility

Category 4: Location-Specific (6 queries)

Travelers who know exactly where they want to stay.

“San Juan Islands vacation rentals waterfront” — island destination; high-intent, geographically narrow
“Mount Rainier cabin rentals” — national park adjacency; tests whether park-adjacent properties appear in AI answers
“Cannon Beach Oregon vacation rentals” — iconic destination; tests whether Haystack Rock-adjacent properties dominate AI citations
“Leavenworth Washington cabin rentals winter” — seasonal + location; tests winter-specific property visibility
“Bend Oregon vacation rentals with pool” — high-growth destination; amenity + location combination
“Whidbey Island vacation rentals” — island destination with ferry access; tests whether accessibility factors appear in AI answers

Category 5: Activity-Specific and Unique Stays (4 queries)

Travelers planning around specific experiences.

“Hood River Oregon vacation rentals windsurfing” — activity-driven; tests whether activity-adjacent properties surface
“Columbia River Gorge vacation homes” — scenic destination; tests whether AI engines distinguish gorge-adjacent from Portland-adjacent
“PNW treehouse rentals unique stays” — unique accommodation type; growing search trend
“PNW group vacation rentals large groups” — group travel; tests whether capacity-filtered results appear

The 6-Engine Testing Protocol

Each query is run fresh on all six engines within a single session window (ideally the same day) to lock the baseline. For engines that support incognito or logged-out sessions, use those to minimize personalization. Record results immediately — AI answers can shift within hours.

Per-Engine Recording Fields

For each query on each engine, record these six data points:

Field	What to Record
Citation (Y/N)	Does the brand or property appear by name?
Position	If cited, what position in the response? (1st mention, 2nd, last, etc.)
Snippet/Surrounding Text	The exact sentence or paragraph where the brand appears.
Competitors Cited	Which other brands or properties appear in the same answer?
Answer Type	Recommendation, comparison, list, or single-entity description.
Source Attribution	Did the engine provide a source link? If so, to what domain?

Engine-by-Engine Notes

ChatGPT (free tier, logged-out session). ChatGPT tends to synthesize rather than list. Record whether it names specific properties, directs to platforms (Airbnb, Vrbo), or gives generic regional advice. ChatGPT’s training cutoff means property availability may be dated; what matters for the baseline is whether it names any real brand at all.

Perplexity. The strongest engine for citation tracking because it provides inline source links. Record every source domain. Perplexity often returns 5-8 sources per query; note which ones are direct property manager sites vs aggregators (Booking.com, Airbnb, TripAdvisor).

Google AI Overviews. These appear above organic results for informational and some commercial queries. Record whether an AI Overview appears at all (presence is itself a baseline signal), which domains it cites, and whether those domains are OTAs, property managers, or editorial (travel blogs, magazines).

Gemini. Google’s model often pulls from Google Maps, Google Business Profile, and travel features. Record whether it returns structured listings (address, rating, price range) or prose recommendations. If structured listings appear, note which properties populate them.

Claude. Claude often declines to give real-time recommendations citing its knowledge cutoff, but may still provide category-level or brand-level mentions. Record whether it refuses, gives generic advice, or cites specific brands.

Meta AI. Accessible via the Meta AI interface. Tends toward conversational recommendations with less source transparency than Perplexity. Record what it cites and whether recommendations appear personalized or generic.

Recording Template

Use this table format to record results. Copy it into a spreadsheet or markdown document and fill one row per query per engine (120 rows total for a full 20x6 baseline).

Query # | Query Text | Engine | Brand/Property Cited? | Position | Snippet (exact text) | Competitors Cited | Answer Type | Source Link? | Notes

Example Filled Row

5 | Oregon coast vacation rentals with ocean view | Perplexity | No (OTA-only: Airbnb, Vrbo) | N/A | "Based on search results, several platforms list Oregon coast rentals..." | Airbnb, Vrbo, Vacasa | List/synthesis | airbnb.com, vrbo.com, vacasa.com | Zero direct property manager citations on Day 0.

What the Baseline Tells You

A Day 0 baseline across 20 queries and 6 engines produces 120 data points. The immediate outputs are:

Citation Rate. What percentage of queries produce any direct brand or property mention? On most first baselines for vacation rental brands, this number is close to zero. AI engines default to aggregator platforms (Airbnb, Vrbo, Booking.com) and major travel publishers (Condé Nast, Travel + Leisure) because those domains have the strongest entity signals and backlink profiles.

Engine Distribution. Which engines cite you and which ignore you entirely? Often a brand appears on Perplexity or Google AI Overviews before ChatGPT or Claude, because the former pull from real-time search indices while the latter rely more on training data.

Competitor Presence. Who shows up instead of you? The competitor column is the most actionable part of the baseline because it shows exactly whose content, schema, and authority signals the engines currently prefer.

Snippet Quality. When you do get cited, what does the engine say? A citation that reads “also available on Airbnb” is very different from “recommended by Condé Nast Traveler as one of the top 10 Oregon coast rentals.”

From Baseline to Proof Cycle

The baseline is Day 0. The same 20 queries should be re-tested at 30, 60, and 90 days after implementing GEO work: structured data markup (FAQPage, Product, LocalBusiness), question-answering content targeting these exact query patterns, authority signals (directory listings, editorial mentions), and technical hygiene. The methodology is identical to the full 50-100 prompt proof cycle described in our measurement guide — just scoped to a single vertical’s query set.

For an aggregated view of what 90 days of GEO typically produces across industries, see What 90 Days of GEO Actually Produces. For the full measurement protocol including matrix construction, gate timing, and client proof package structure, see How to Measure and Prove GEO Results.

Limitations and Honest Notes

Programmatic search engine access is restricted. Google, DuckDuckGo, and Bing block automated queries; the AI engines (ChatGPT, Perplexity, Gemini, Claude, Meta AI) do not expose public search APIs for fresh-session citation testing. A proper baseline requires manual execution: a human opening each engine interface, entering each query, and recording results in the template above.

This post provides the framework, the query set, and the recording protocol. The execution — the actual manual querying and recording — is the operator’s responsibility. What Stay Citable delivers for clients is exactly this manual execution at scale (50-100 prompts across 6 engines), with the matrices compiled, analyzed, and delivered as a prioritized roadmap.

Automated alternatives exist for specific engines: Perplexity offers an API (requires approval), and Google’s Search Console and Programmable Search Engine provide partial signals. But none of these replicate the fresh-session, logged-out query experience that real users get — and that is the experience that matters for citation visibility.

FAQ

Why 20 queries instead of the full 50-100?

The full 50-100 prompt matrix is appropriate for an ongoing client engagement where the goal is comprehensive share-of-voice tracking. A 20-query vertical baseline is a lighter-weight starting point that still produces statistically useful signal: 20 queries x 6 engines = 120 data points, enough to identify patterns, gaps, and the highest-leverage first moves without the overhead of a full matrix.

How often should the baseline be re-tested?

For seasonal businesses like vacation rentals, re-test quarterly to capture seasonal search pattern shifts (summer beach queries vs winter ski/mountain queries). For year-round categories, 30/60/90-day gates are standard. The query set itself should be reviewed every 6 months as traveler search behavior evolves.

What if an engine doesn’t return results for my queries?

Record that as data. If Claude refuses to recommend specific properties or Google AI Overviews don’t trigger for a query, that’s a signal about how that engine handles your category. Some engines are inherently better targets for certain verticals. The baseline reveals that.

How do I know if my baseline numbers are good or bad?

The baseline is not a score — it’s a starting point. A 0% citation rate on Day 0 is normal for brands that have never done GEO work. What matters is the delta at Day 30, 60, and 90. The baseline exists to make those deltas measurable and defensible.

Can I use this framework for a different vertical?

Yes. The 20-query structure, engine protocol, and recording template are vertical-agnostic. Replace the PNW vacation rental queries with your own category’s high-intent search terms, organized into the same five intent categories (discovery, amenity, budget, location, activity/unique). The engine protocol and recording fields remain identical.

Does the free audit cover this?

Yes. The free citation audit delivers a Day 0 baseline on your actual prompt matrix plus a prioritized 60-90 day roadmap. For vacation rental brands, we would work with you to identify the 50-100 most important queries (including variations on the 20 above) and run the full matrix across all six engines. See our Free AI Citation Audit Checklist or visit the audit page to start.

Next Step: Lock Your Day 0 Baseline

If you manage vacation rental properties in the Pacific Northwest — or any competitive travel market — the single highest-leverage thing you can do this week is lock a Day 0 baseline. Without it, any visibility improvements later can’t be proven or attributed. With it, you have a defensible starting point for measurable GEO.

Get your free citation audit. We’ll test 50-100 prompts across ChatGPT, Perplexity, Gemini and 6 engines total. Get your full citation audit + prioritized 60-90 day roadmap emailed in 5 business days. No credit card. No sales call.

Get your free citation audit →

Business Impact

For travel and hospitality brands, AI citation visibility directly affects booking volume. A 35-55% citation-rate shift — the typical Day-90 outcome across our 50-100 prompt, 6-engine client matrices — commonly produces 150-400% lift in tracked AI referral sessions. In the vacation rental vertical specifically, where booking values are high and purchase cycles are short, even a 10-15% citation gain on high-intent location queries (Cannon Beach, San Juan Islands, Mount Rainier) can shift tens of thousands of dollars in revenue per month away from OTA commissions and toward direct bookings. The Semrush January 2026 benchmark showing 15.9% AI-referral conversion (versus 1.76% Google organic — a 9x differential) makes the case: AI citations convert. A Day 0 baseline is the first step to capturing that conversion advantage.

Sources

Previsible LLM referral session study (1.96 million sessions tracked), 2025
Semrush AI referral conversion benchmark, January 2026
Princeton GEO study (Aggarwal et al. KDD 2024) — up to 40% citation improvement from structured signals
Stay Citable Day 0-90 proof cycle methodology and aggregated client results, 2025-2026
Client matrices and case studies (SaaS, professional services, ecommerce verticals), 2025-2026
Google AI Overviews trigger behavior observed across commercial and informational travel queries, 2025-2026
Perplexity source attribution patterns for travel and hospitality queries, 2025-2026

See also the vertical-specific case studies: B2B SaaS: 42% Citation Rate, Professional Services: 51% Citation Rate, and What 90 Days of GEO Actually Produces.