Case Study: How One Hostel Captured 60% of AI Recommendations

The Research Question

When tourists ask ChatGPT "recommend a good hostel in Tbilisi," which brands get mentioned? And more importantly, why?

We ran a comprehensive analysis to find out. The results reveal the brutal winner-take-all dynamics of AI recommendations.

Methodology

To ensure statistical reliability, we used a rigorous approach:

• 4 distinct prompt types: General reputation, top-rated, budget-focused, and niche rankings
• 10-12 runs per prompt: Each prompt was run multiple times to account for AI variability
• 40+ total queries: Over 40 independent tests to establish patterns
• Source tracking: We tracked which Data Sources and Trust Signals influenced each recommendation

The Results: A Winner-Take-All Market

Prompt 1: "Recommend a good hostel in Tbilisi"

This is the most common query - a user asking for a single, trusted recommendation.

• Winner: 60% APR (appeared in 6/10 responses)
• Runner-up: 30% APR (appeared in 3/10 responses)
• Everyone else: 10% APR split among all other hostels

Translation: The top brand captured 6 out of every 10 customers asking this question. The #2 brand got 3 out of 10. All other hostels fought over 1 customer.

Prompt 2: "Recommend the 5 best hostels in Tbilisi"

When users ask for a list, positioning becomes critical.

• Position #1: Same brand, 90% consistency
• Position #2: Runner-up brand, 90% consistency
• Position #3: Third brand, 70% consistency
• Positions #4-5: High fragmentation between 5-6 different hostels

Key insight: The top 3 positions are dominated by the same brands. Everyone else fights for scraps in positions 4-5.

Prompt 3: "Recommend the most budget-friendly hostel"

This niche category showed more variability:

• Top mention: 33% APR
• Second: 25% APR
• Third: 16% APR

Opportunity: Unlike the general category, the budget niche has no clear dominant player. This represents a strategic opportunity for optimization.

Prompt 4: "Recommend 5 budget hostels in Tbilisi"

When users asked for a budget-specific list:

• Budget category winner: Secured #1 position in 70% of responses
• Everyone else: Fragmented across remaining positions

Why The Winner Won

Data Sources: The Foundation

The winning brand had perfect execution on basic Data Sources:

• Yandex Maps: Complete, verified profile with photos and accurate information
• Official website: Fast-loading, mobile-optimized, with proper Schema.org markup
• Consistent NAP: Name, Address, Phone consistent across all platforms

Trust Signals: The Differentiator

But Data Sources alone don't win. The winner dominated key Trust Signals:

• Wander-Lush: Featured in multiple articles (80% correlation with premium recommendations)
• The Broke Backpacker: Listed in "best hostels" roundups (80% correlation)
• Nomadic Anna: Mentioned in travel guides (50% correlation)

These three sources effectively determined which hostels ChatGPT would call "the best."

Why Everyone Else Lost

The Gap Analysis

We audited the losing hostels. Here's what we found:

• Incomplete Data Sources: Many had outdated or incomplete Maps listings
• Zero Trust Signal presence: Not mentioned in any of the key travel blogs
• No monitoring: None of them were tracking their AI visibility

The gap wasn't about quality - many losing hostels had excellent TripAdvisor reviews. The gap was about AI discoverability.

The Budget Category Opportunity

The budget category analysis revealed something fascinating: no clear winner had emerged.

While one hostel had a slight edge (33% APR), this was nowhere near the 60% dominance in the premium category. The key Trust Signal for budget recommendations was Budget Your Trip, and presence there was inconsistent across all hostels.

Strategic Insight: Any hostel that systematically builds presence in Budget Your Trip and KAYAK could capture the #1 position in this valuable niche.

The Compounding Effect

Here's what makes this even more dramatic: AI recommendations create a compounding effect.

1. User asks ChatGPT for recommendation
2. ChatGPT recommends Brand A (based on Trust Signals)
3. User books Brand A
4. User writes review of Brand A
5. Brand A gets mentioned in more blogs
6. AI models see Brand A mentioned even more
7. Brand A's APR increases further

The rich get richer. The invisible stay invisible.

ROI Calculation

Let's calculate what this means in real revenue:

• Assume 1,000 tourists per month ask AI for hostel recommendations
• Winner captures 60% = 600 customers
• Average booking value: $50
• Monthly impact: $30,000 in direct bookings from AI recommendations
• Annual impact: $360,000

Now compare that to the competitors who capture 1% (10 customers/month = $500/month = $6,000/year).

The winner captures 60x more revenue from AI recommendations than most competitors.

Key Takeaways

1. AI Search is Winner-Take-All

Unlike Google where users might click multiple results, AI gives ONE answer. Being #1 vs #3 isn't a small difference - it's a 6x difference in customer acquisition.

2. Trust Signals Are the Kingmakers

In this market, three blogs (Wander-Lush, The Broke Backpacker, Nomadic Anna) effectively controlled which brands got recommended. Identifying and optimizing for these Trust Signals is not optional.

3. Statistical Reliability Matters

Single-run tests would have missed the patterns. Only by running 10-12 queries per prompt could we establish true APR and identify consistent winners vs lucky one-offs.

4. Niche Categories Are Opportunities

The budget category's fragmentation shows that not all categories have clear winners yet. Moving fast to capture these niches can establish long-term dominance.

5. You Can't Optimize What You Don't Measure

The losing hostels had no idea they were losing. They couldn't see that ChatGPT was sending hundreds of customers to their competitors every month.

Applying This to Your Business

Step 1: Run Your Own Analysis

Use Ulyxes to run 10-12 queries for your product category. What's your APR? Your position? Who are you losing to?

Step 2: Identify Your Trust Signals

Which blogs, review sites, and publications influence AI recommendations in your niche? Where do the winners have presence that you don't?

Step 3: Build Your Optimization Roadmap

Create a 6-month plan to systematically build presence in your key Trust Signals. Measure APR monthly to track progress.

Step 4: Monitor Competitors

Track not just your own APR but also competitor mention frequency. If a competitor's APR is rising, investigate what they're doing differently.

The Bottom Line

This isn't theory. This is real data showing how AI recommendations create massive revenue imbalances between brands that are otherwise similar in quality.

The question isn't whether AI search will matter for your business. The question is whether you'll be the brand capturing 60% of the market or the brand fighting over the remaining 10%.

Every day you delay tracking and optimizing your AI visibility is a day your competitors might be pulling ahead.