Explanation runs through much of this question. How we interpret metrics, markets, consumers or frameworks determines what we build, how we compete and what we believe is true.
special thanks Propelik Sponsor this newsletter:
NavLog is a research-focused biweekly newsletter that turns real travel data into strategies you can actually use. Are you working on a budget? This is worth reading. Less noise, more signal, zero versatility → Join over 2,000 travel marketers
Both Amazon and eDreams can proudly display the Prime chart that pops up to the upper right. In both cases, Prime membership growth has been impressive. The difference becomes apparent when you plot total revenue on the same graph.
For eDreams, Prime membership grew approximately 14-fold, from approximately 500,000 to nearly 8 million. Revenue margin, eDreams' preferred revenue metric, grew only slightly over the same period.
Amazon's early Prime years tell a different story. As Prime membership has grown from millions to around 100 million, the company's total revenue has also risen significantly.
Premium membership should be cause for celebration, but it should also drive the core business forward. eDreams has built an impressive subscriber base. But nine years later, revenue indicators tell a different story.
Air Transat is not a World Cup sponsor, but all they need is a billboard, two numbers and a pain point. For the price of 90 minutes at the stadium, you can spend a week in the country you support. Of course, some fans will see this and spend $429 to fly to Mexico and then spend $3,563 on a World Cup ticket.
One purchase leads to another purchase. In 1769, Denis Diderot received a new dressing gown and ended up replacing everything in his home to match it, depleting his savings in the process.
The journey begins. You book a flight and suddenly your luggage doesn't look like enough. Book a hotel and suddenly a better room is only $40 more. To book a tour, you need gear. The entire travel upsell industry is built on the fact that one purchase makes the next purchase feel necessary.
When travelers ask ChatGPT where to stay in Barcelona, most hotels in the city don't come up. Lighthouse ran 4,545 tips across nine destinations and five traveler personas and found that AI recommendations were highly concentrated; in Tokyo, only 10% of hotels were mentioned. In Paris, it's 13%. In a market of approximately 2,000 hotels, one hotel in Paris accounted for 3% of all mentions.
Independent hotels are underperforming. In Indianapolis, they received 6.5% of mentions, even though they accounted for a much larger share of supply. AI recommendations will also impact luxury goods. Even with generic tips and no budget signals, four- and five-star hotels dominate. There are almost no three-star hotels.
Two findings stand out. First, 82% of AI sources come from OTAs and editorial sites like Forbes, Lonely Planet and Condé Nast. Getting on “best of” lists is now a clear distribution strategy. Second, when AI recommends hotels and travelers click, most clicks go directly to the hotel website, bypassing the OTA entirely. Reading + Lighthouse
Mark Pincus built a framework at Zynga called Proven Better New. Isolate your one true innovation and copy all the other products that already work on a pixel level for your platform and audience. When Sid Meier (one of the most influential game designers in history) launched a Social Civilization game on Facebook, Zynga's product managers declared the game dead within an hour, and they turned out to be right. He didn’t copy their onboarding process, which had proven to be effective.
The problem is that founders become founders precisely because they want to innovate. Plagiarism feels like cheating. Pincus calls it ethical arbitrage: the instincts that make you an entrepreneur can lead you to over-innovate and lose users before they get to your actual idea.
This goes against the common instinct that founders are often told to focus internally and worry less about competitors. The logic is that most startups die by suicide, not murder. This is compounded by the fact that new products are proving to be better. You need a deep understanding of the competitive landscape to know exactly which single element you're going to bet on.
Another reframe is who you are building for. If you define your ambitions in the eyes of your peers, you will over-innovate. If you define it from the perspective of your consumer (the nurse in Indiana playing Farmville), you copy something she already likes, make it an inch better, and add one more thing she didn't know she wanted. As Pincus says, burning your resume means defining ambition in the eyes of consumers rather than peers.
Mark Pincus, founder of Zynga and one of the most prolific consumer product manufacturers in Silicon Valley history, on Lenny's podcast Recently, he cited travel agencies as one of the first examples of latent consumer needs that artificial intelligence could eventually unlock. He believes that the reason why we have not found an excellent travel agency that is always online is because economic conditions do not allow it. The commission model cannot support 24/7 personalized service for each traveler, and consumers are not willing to pay enough for this. The Internet solved the price problem, but not the service problem.
What he describes is understanding your travel environment, being available 24/7, and being ready to take action if things go wrong during the trip, such as rebooking flights, managing logistics failures in real time, etc. That last part is where the value lies and where current travel apps fall short.
He also raised the question of whether OpenAI or Anthropic would build this themselves. Travel agents, he says, are exactly one consumer use case that could give people a clear reason to choose one LLM over another. Currently most users cannot differentiate between Claude and GPT. They can solve this problem through consumer use cases such as travel.
While Booking, Airbnb, and Hilton don't appear in ChatGPT ads, Expedia has no rivals in accommodations. Flavio Rodrigues (Digital Sardine) has been within the platform for three weeks and found no other major competitors among the dozens of queries he tested. He also found that Expedia wasn't doing anything new. They took their existing Dynamic Search Ads infrastructure, Destination × Property Type × Amenities, and applied it to new channels.
Entry point mathematics explains part of the gap. Minimum CPC is $3. At a 2% conversion rate, each booking costs approximately $150. Most OTAs can't make the numbers work. Expedia can do just that with its bundled add-on economy. The remaining gaps are operational lag, channel skepticism, and in Airbnb's case, deliberately chosen to avoid performance marketing.
Flavio's post details how the ChatGPT platform works, including targeting mechanics, bidding floors, who will see ads, and what travel brands need to know before testing. Read+
Lennart Dobrowski wrote LinkedIn Post That goes deep Expedia’s AI Trust Gap Reportfound that 66% of travelers don’t trust AI to book travel on their behalf. Skift, which calls itself “the trusted source of intelligence on the global travel industry,” cited the Expedia report with the headline “Expedia: Only 8% trust AI to book travel.“But where did that 8% come from? As Lennart mentioned, if 66% of people don't trust AI to book on their behalf, it's fair to assume that the remaining 34% either believe in it or are at least open to the idea. Lennart reviewed the report and found a statistic that says 8% of travelers currently rely on artificial intelligence platforms for travel planning. But using artificial intelligence for travel planning and trusting artificial intelligence to book travel are very different things. However, the headline became “industry truth” within 48 hours.
This was in April. Two months later, I asked Cloud what percent of travelers trusted AI to book travel. Crowder confidently cited the Expedia report, which pointed to 8%. When I mentioned that Expedia meant usage, not trust, Claude himself understood: “you're right. This is using statistics, not trusting statistics. The 8% figure also appears in Skift’s headline – “Only 8% trust AI to book travel” – which treats this as a trust statistic. The title is a misrepresentation of the original data and I repeat it uncritically.“
Lennart's main point is correct: always read the chart, then read the issues behind the chart. I would also add to focus on the motivations of the person telling the story. “Only 8% trust AI to book travel” leads to more clicks than “8% rely on artificial intelligence for travel planningTwo months later, the headline and article remain the same, and the explanation has now spread far enough that the artificial intelligence system repeats it as fact.
I use it all the time to stress test early stage ideas and think about go-to-market issues. I got harsh feedback and angles I hadn't even considered. — Emma Burton, Head of Operations and People, Oliver's Travels
this Travel Technology Essentials Co-Pilot Currently has over 1,000 conversations and a rating of 4.5. Give it a try and let me know what you think.
→ Explore All 1622 vacancies Currently on the Tourism Technology Essentials Working Committee.
📩 For monthly updates on the latest roles, please subscribe Travel Technology Work Newsletter
If you're a startup looking to raise a round of funding (from pre-seed to Series D), I can help (for free). Tourism Investor Network is a private platform where I recommend innovative travel startups to investors and innovators. If you are interested, please start from fill in this form.
If you find value in Travel Tech Essentialist, the best way to express it is by sharing it with others who appreciate it:
If you haven't subscribed yet, please join us here:
Thank you for trusting my inbox.
Mauricio Prieto
