A flight search product that can't get the data. A bankrupt airline whose operating history is worth $10 million. Loyalty programs running out of seats to give away. Banks paying travel brands more at every renewal.
Special thanks to Propellic for sponsoring this edition of the newsletter:
Your audience is traveling more than ever, so why are bookings slipping?
NavLog, the biweekly digest by Propellic, digs into questions like that to help you turn real data into revenue-driving strategy. Recent issues unpack why demand and bookings can move in opposite directions, and how to stop your multi-domain portfolio from outbidding itself. Join 2,000+ travel marketers
Ordinary abundance
One of the best things I’ve read this year. It will change how you appreciate the world. Jordan Dworkin built a simple site that walks you through a modern apartment and pairs each object with someone from history who found it unimaginable. Electric light, cold storage, clean water from a tap, eyeglasses, etc…
New York to London now takes seven hours and costs a few hundred dollars. But most of the complaints are about the Wi-Fi and the boarding process.
Selling the search
Ed Fry is the friend everyone sends their flight problems to. He has flown about a million miles and will happily spend an evening finding someone a better routing. Two years ago he started building a product to do that automatically for UK travelers, turning that evening into two minutes. The product works, but he can’t launch it.
His write-up explains why consumer flight search has barely changed since 1999. Airlines cap how many searches a partner can run for each booking it produces, because searches cost them money while bookings generate revenue. Metasearch sites and travel agents all work within that cap. Aggregators keep a stored copy of prices rather than asking the airline fresh each time, which is why the fare sometimes jumps when you reach checkout.
Fry couldn’t get a licensed feed, so he built around it. Great British Take Off works out which searches are worth running for a given trip, runs them, and then sends you to the airline or the OTA to book. £3.20 per trip or £89.64 a year, which is the 1 to 2% of airfare he argues good search advice is worth on its own.
He charges for the searching and leaves the booking to whoever holds the data.
Spirit’s OS
Google is paying $10 million for a de-identified archive of Spirit Airlines’ internal operations: 100 million emails, 500 million Teams records, millions of files, 30 million lines of code, and billions of records covering pricing, reservations and disrupted operations. The interesting part is how all these interconnect. Google can potentially see an IT ticket, the emails discussing the problem, the code change, the review comments, the operational result, and what happened to the business afterward. It’s a record of how a complex company made decisions. That’s very different from training an AI model on another billion webpages.
We tend to think of proprietary data as customer profiles, searches and bookings. But in the AI era, one of the most valuable assets may be the accumulated record of how your company operates and makes decisions. Spirit went bankrupt, but decades of operating knowledge still turned out to be worth $10 million.
A failed airline is becoming training material for the future of enterprise AI. The failure could be making the dataset even more useful, as it contains not only what worked, but what didn’t, and the consequences.
The seats that used to be rewards
Deloitte surveyed 5,564 US loyalty program members, including 683 airline members. Some flyers give one airline every dollar they spend on flights. They are the most loyal customers an airline has, and the least convinced the program is worth anything. Only 23% say it adds value, compared with 32% of flyers who give the airline 75 to 99% of their spend.
Deloitte thinks part of the explanation is that airlines got better at their jobs. Load factors went from 72% in 2000 to 82% in 2025, and premium cabins increasingly sell rather than sit empty. Award seats and free upgrades were what loyalty programs had to give away, and they came from seats nobody had bought. There are fewer of those seats now. Fifteen years of unbundling also turned many of the other perks into things you can simply pay for.
There’s a generational problem too. Gen Z and millennials are half of all travelers but only 27% of loyal members. The assumption was that younger travelers would concentrate their flying as they got older. Millennials are 30 to 45 now and still look closer to Gen Z than to Gen X.
Continue on Google
Google added three things to AI Mode in Search yesterday. Flight price tracking, now covering more than 180 countries. Points and miles pricing for flights and hotels, starting with American, Alaska/Hawaiian, Choice, Hilton and Wyndham, with Accor, Flying Blue, Hyatt, LATAM and Lufthansa coming soon. And hotel booking you can finish without leaving the chat, rolling out in the US now.
To complete a booking, you pick a hotel, tap “Continue on Google” next to one of the integrated partners, choose a room, and pay with Google Pay. The hotel or platform is still the merchant of record and still handles customer service if something goes wrong. Booking.com, Choice, Expedia, Hilton, Hotels.com, IHG, Marriott, Priceline, Trip.com and Wyndham are in at launch. They supply the rooms, keep the payment relationship, and take the calls. Google just owns the screen where you decide.
Google tried this once already. “Book on Google” let people book a hotel straight from search results back in 2015, and Google killed it in May 2022 for low take-up. Now hotels are the first vertical back for agentic checkout; same idea, new interface.
Comparing award value across loyalty programs used to take real effort. Once AI Mode can put miles pricing for five or more programs side by side on request, that effort disappears. And for programs whose heaviest users already doubt they’re worth much (see above), removing effort isn't the kind of transparency that helps the case for holding onto the points.
The balance of power is shifting from banks to travel brands
Morgan Stanley puts co-branded travel card revenue at about $24 billion a year today and projects it could reach $60 billion by 2035 in its base case, or $100 billion in the bull case. Premium travel cards have grown roughly 10% a year since 2019, versus 6% for cards overall.
Terms keep moving in favor of the travel brands. Royalty fees paid by banks to airline partners have been growing at low double-digit rates annually. American generated $6.2 billion from co-brand and partner agreements in 2025. Delta collected $8.2 billion from American Express and expects that figure to reach $10 billion within the next few years. Alaska’s co-brand remuneration grew 10% last year.
As consumers demand richer rewards and travel brands negotiate higher payments, banks are getting squeezed from both sides. The credit-card pie is getting much bigger, but travel brands may be taking a larger slice. Read + TheStreet.
Sleeping at the airport is a feature request
Donna McSherry started Sleeping in Airports in 1996 as a one-page review of Dublin, Frankfurt, and Geneva. It now covers more than 800 airports.
On July 22 (National Hammock Day), TSA (the US Transportation Security Administration) posted photos of travelers who had strung up hammocks to sleep through a layover. They got 1.1M views. JoseLuis Vilar commented that he once tried to pay for a nap pod on a US layover, the card reader didn’t work, and he went back to the metal chairs like everyone else. His takeaway was that every improvised solution is your customers telling you what to build, and for free.
Some airports worked this out years ago and built rest areas and quiet zones. Others are looking at people asleep in hammocks and seeing a photo op.
The less glamorous AI opportunity in travel
PhocusWire asked investors and operators what they’re looking for in the next generation of AI travel startups. I contributed to the piece. Read + PhocusWire
My take is that one of the biggest AI opportunities in travel is in the non-glamorous stuff: workflow automation. There’s a lot of focus on AI agents that plan and book trips. Meanwhile, much of the travel industry still runs on people handling schedule changes, processing refunds, reconciling suppliers, assisting call centers and managing itinerary changes. Automating this work isn’t as glamorous as building the next AI trip planner, but the problem is painful, the ROI is measurable and customers already spend a lot of money solving it.
The travel VCs I speak with have also become wary of “AI-native” as a pitch in itself. A better model isn’t much of a moat when everyone has access to better models. Supplier relationships, proprietary data, operational know-how, distribution and customer trust are much harder to replicate.
Raising a round?
If you are a startup looking to raise a round (from pre-seed to Series D), I can help (for free). Travel Investor Network is a private platform where I recommend innovative travel startups to investors and innovators. If you’re interested, please start by completing this form.
Travel Tech Essentialist Job Board
→ Explore all 1431 open roles on the Travel Tech Essentialist Job Board now.
Expedia | Director of Technology | Seattle | $201k - $282k
Fora | Senior Lifecycle Marketer | New York City | $160k - $220k
The Hotels Network | Regional Sales Manager - LATAM | Denver | $102k - $130k + Equity
Airbnb | Global Community Programs Lead | $156k - $195k + Equity
📩 For monthly updates on the latest roles, subscribe to the Travel Tech Jobs newsletter
🧭 Travel Tech Essentialist Copilot
Tested it out and it’s given some really great feedback — making changes accordingly. — Jared Lerner, Founder & CEO, Nappr
The Travel Tech Essentialist Copilot has now had over 1,000 conversations and holds a 4.5 rating. Try it with one problem you’re wrestling with and let me know where it gets things right or wrong.
If you’ve found value in Travel Tech Essentialist, the best way to say so is to share it with someone who’d appreciate it:
If you’re not yet subscribed, join us here:
Thanks for trusting me with your inbox.
Mauricio Prieto
The Spirit example made me think about something beyond proprietary data. An archive becomes valuable not just because it contains many records, but because it keeps the links between a problem, the discussion, the decision, and the result. That is more useful than a customer profile alone. What context should remain before a travel company lets an AI system treat an old operational record as useful knowledge?
Thanks for all your research. I find it American heavy, and honestly, I'm no longer interested in American information. They are fast becoming a banana republic. I would be more interested in Global aviation research going forward.