104-Punch and Counterpunch in the Online Airline Industry

Seventeen years after the original blog below, two of its three protagonists have vanished into a single corporate parent, the fee fight it describes has been replaced by a hotel-commission fight nobody sees, and the company most responsible for reshaping how people book a trip wasn’t even founded when this blog first ran. This update covers what happened between 2009 and now: a wave of acquisitions because buying  was cheaper than winning, a Leader’s Trap that outlasted the price cut that supposedly ended it, and a concentration in this industry so complete that the today’s description isn’t a duopoly — it’s four companies holding ninety-six cents of every revenue dollar.

Posted 5/14/09

Orbitz Worldwide, the online travel booking agency, competes with the likes of Travelocity and Expedia. Of the three, Orbitz is the most dependent on airline booking fee revenue for its profits. Travelocity and Expedia both reduced fees for their booking of airline tickets before Orbitz. Orbitz held on to protect its margins. Orbitz began losing market share and reversed course. It announced that it would waive booking fees on most flights booked through May. This brings its pricing in line with its competition.

Leader’s Trap Examples – StrategyStreet.com

Then Orbitz did the industry one better. In an effort to grab market share and punish its discounting competition, the company announced a new promotion called “Dare to Compare.” This program brings with it a reduction in service fees on hotel rooms booked on its web site through July 15th. Orbitz is hoping to gain enough market share to offset the reduction in its fees.

Normally, an industry with only three major players is able to protect its pricing structure. Usually, the three players decide there is little to gain in price competition with one another. Apparently, this industry thinks differently. Orbitz most recent reporting found revenue off by 14%, as travelers cut back in the tough economy.

These industry discounts are a waste of margin. The entire industry is likely to copy any leading competitor’s price discounts. After all, the industry lives with the airline industry, where minute-by-minute price matching has become an iron-clad rule. These discounts don’t mean much to consumers so they are unlikely to energize demand.

Orbitz also took another step recently that is far more promising. It launched a price-assurance program that automatically refunds customers when a hotel price drops below the one they paid appears on its web site and is purchased by another customer. This innovation improves the company’s Reliability in customer eyes (see the Perspective, “Discovering Hidden Pricing Power” on StrategyStreet.com).

For many ideas to improve market share and profits by a judicious use of pricing, see www.StrategyStreet.com/Improve/Pricing/Brainstorming. Then follow the choices

***

Update 2022:

The online travel agency business has changed a great deal since 2009. It is no longer as competitive as it once was. Over the years, Expedia has purchased both Orbitz and Travelocity. In 2021 the combination of Expedia and Booking (formerly Priceline) controlled a large percentage of the market.

The sources of revenue have also changed drastically. Airlines are no longer so important. They generated only 3% of Expedia’s revenues in 2021. 75% of its revenue came from lodging. Expedia typically charges commissions of 15 to 30% of the revenues it produces. About 53% of this revenue comes using a merchant model where Expedia purchases hotel rooms in bulk and then resells them to the market. The other 47% comes from an agency model, where Expedia does not own the hotel rooms but simply brings traffic to the hotel company in return for its commissions.

The industry leaders of 2022 have made effective use of acquisitions to achieve their market dominance. See HERE and HERE for some thoughts about the effective use of acquisitions.

***

Posted 9/26

This Blog was not about one business — it was about two, welded together by habit. Business 1 (Pattern: a group of customers who buy defined products from a set of competitors; the business changes if any two of these change)  rents locations to consumers: hotels, chains, and the millions of individual hosts who actually own the rooms. Business 2 sells nothing but the search, the comparison, and the click that gets you to one of them and takes a commission for it. Every company this blog has ever named — Expedia, Orbitz, Travelocity, Booking, even Airbnb — has always been a Business 2 company; not one of them owns a bed. This update covers Business 2 only.

 

Booking Holdings and Expedia Group both make good money running Business 2. Booking makes noticeably more of it, and the gap is real — but it is not a story about one company beating the other. It is a story about a market that has quietly become vulnerable to its own biggest customers, and about a rival that has already tried to blow the whole thing up once and backed away.

Two Businesses, Not One

Recall the opening of this post, the three-way fee fight it describes, and it reads today like a category error. Orbitz, Travelocity, and Expedia were not competing to rent anyone a room or a seat — they were competing to be the search box a traveler clicked through on the way to one. That is Business 2: a platform that enables the online purchase of travel inventory it does not itself own. Business 1 is the inventory itself — hotels, chains, and the individual hosts who list on Airbnb — and it has its own competitors, its own economics, and its own separate strategy demands. The SEC’s classification agrees with the split: Expedia, Booking, and even Airbnb are all classified under services and travel-agency codes, not hotel or accommodation codes. Airbnb owns no beds. Neither does Booking. This update stays entirely inside Business 2.

Business 2, measured the right way — by its own commission and fee revenue, not by the value of the trips booked through it — took in roughly $58 billion in 2024, with Booking, Expedia, Airbnb, and Trip.com together collecting about 96% of it. That is a Stable Market (Pattern: an industry with low positive price volatility and attractive average returns on investment): returns comfortably above Hostile-Market territory at both major players, no Price Umbrella (Pattern: when incumbents hold prices up, letting others discount against them and take share) held by one Business 2 competitor against another, and no company-wide discount war since the one this post opened with.

Orbitz’s Discount Was a Symptom, Not a Cure

The original fee fight still ended exactly the way it looked like it would. Orbitz held its price longer than its rivals, broke late with an underpriced “Dare to Compare” campaign, and spent six more years as the weakest of the three protagonists before Expedia bought it in 2015. That is a clean Leader’s Trap (Pattern: when an incumbent holds a price umbrella and loses share to a discounter; the price fight itself teaches customers distrust, turning a pricing failure into a lasting Reliability failure that outlives the eventual price cut) — the pricing failure came first, but it was the distrust that followed that kept Orbitz weak long after its price was back in line. It became an asset, folded into Expedia’s operation rather than shut down — the same math behind every acquisition in this update: a competitor losing share at even half the industry rate takes roughly forty years to be fully absorbed organically, and nobody runs a company on that plan.

What Actually Separates Booking From Expedia

Run the two of them through the Customer Buying Hierarchy (Pattern: the consistent order customers use to evaluate alternatives: Function, Reliability, Convenience, then Price) and most of it turns out to be a wash. Function (Pattern: the product characteristics that affect how the customer uses it, essentially all of its features) has converged — comparable apps, comparable real-time inventory, comparable search-and-book mechanics at both. Booking does hold one specific Function edge: roughly 4.4 million properties listed, against a customer base still about 90% independent hotels and small operators rather than chains. That is reach, and reach is Function — more selection, more capability — not proof of anything deeper.

Convenience (Pattern: the ease of acquiring a product, measured by Order Cycle Time, the elapsed time from deciding to buy to beginning use) is also a wash, and worth being clear about: metasearch — Kayak, Trivago, Google Hotel Ads — shortened search-to-book time for the whole industry at once, not for one platform against the other. There is no evidence Booking’s Order Cycle Time beats Expedia’s, or the reverse.

Price, by contrast, has not converged at all, and this is where a story calls for an explanation. The explanation is product mix. Booking’s take rate has run around 16% of a booking for years; Expedia’s closer to 11%. That is not Booking exercising pricing power over a weaker rival — it traces to a genuinely different product mix. Booking’s Europe-heavy, independent-hotel-dominated base commands more price and a different business model. Expedia leans harder on lower-margin prepaid bookings. Two companies selling a nearly identical product at a nearly identical headline commission range are still running two different cost structures underneath, and that gap, not a Function or Convenience advantage, is most of what shows up in the return numbers: a 37% adjusted EBITDA margin at Booking against roughly 24% at Expedia.

That leaves Reliability (Pattern: the consistency with which a company delivers on its promises: reliability of delivery, of function, and of market presence), and this is where Booking’s real edge lives. Genius, its loyalty program, is worth looking at closely because it is not one thing — it is a Function feature (a specific, concrete mechanism: better rates and free upgrades for enrolled bookers) that does its actual work as a Reliability play, keeping repeat travelers coming back because the platform has consistently delivered on what it promised them. Expedia has an equivalent in One Key, but built its scale primarily by acquiring Travelocity and Orbitz rather than by winning that trust directly — which is a perfectly good way to buy volume and a much slower way to buy Reliability and a return.

The Customer Who Doesn’t Pay, and the One Who Does

Business 2 is a two-sided market, and it is worth being precise about which side actually pays for it, because the two sides want different things and Business 2 has to serve both. The traveler chooses which platform to search and book through — that choice is what Booking and Expedia are fighting over with Genius and One Key — but the traveler does not pay Business 2 a fee for the privilege. The hotel does. Commission on each booking is the entire revenue line this update has been discussing, and the hotel paying it is disproportionately a small one: roughly 90% of the properties on Booking’s platform are independent operators, not chains. That is close to backward from the usual pattern, where an industry’s largest customers use their size to squeeze suppliers on price. Here, the largest “customers” — the big chains — have simply started supplying themselves, and the online travel agencies are left more dependent on the medium and small end of their own customer base.

Airbnb and Trip.com belong in this same two-sided frame, not off to the side. Airbnb is a Next Leader (Pattern: a competitor or product offering much better than standard performance at a low price to a specific customer subset, made possible by a low cost structure) purely as a Business 2 company — it owns no rooms, and its entire advantage is a Cost Structure (Pattern: a company’s total costs: operating costs plus capital costs) with no owned inventory behind it, connecting travelers to a Business 1 supply base (individual hosts) nobody else had organized at scale. Trip.com runs the same commission-taking playbook Booking and Expedia do, just built around Asia-Pacific supply and demand instead of Europe and the U.S. All four companies are doing the identical thing — selling the click, not the room — which is exactly why lumping any of them in with Marriott or an independent host was always a category error.

The Price Umbrella Nobody’s Watching

Here is the part of this update I think matters most going forward, and it did not come from the numbers — it came from asking more questions about them. Hotel chains are integrating backward into Business 2, building their own direct-booking apps specifically to stop paying commission. It is tempting to read that as just a competitive threat sitting outside Business 2’s usual roster. I think it is better read as a warning.

There is an implicit rule worth stating plainly, because it explains everything else in this section: your price cannot incite your customer to build its own version of your product. That is not a metaphor here — it is literally what a hotel chain’s direct-booking app is. The moment a supplier’s commission bill outweighs what it would cost that supplier to build the alternative, staying loyal stops being the safe choice and starts being the expensive one.

A Price Umbrella is normally a story about one competitor holding price while a rival undercuts it. What is happening here runs on a different axis, but the same underlying mechanism: Business 2’s commission may simply be priced high enough that building a substitute became the rational move once a supplier could afford to build one. For Marriott, Hilton, and Hyatt, that tipping point has already been crossed — it is too late to reprice against that tranche; they are already gone, or going. The live strategic question is where that same tipping point sits for the next tranche of supplier size down: the large independent groups and mid-size chains who cannot yet justify the build, but might soon, especially if the build itself keeps getting cheaper. If Booking and Expedia are not actively pricing against that specific tipping point, they will find out where it sits the same way they found out with the big chains — after the fact, and one tranche too late.

There is a real lever here, and it is the same one this whole update keeps returning to: the marginal producer, not either company’s own cost structure, sets the effective ceiling on what commission the market will bear. Today that marginal producer is a mid-size chain’s build-versus-buy decision. A hotel weighing whether to build its own booking app is not comparing Booking’s price to Expedia’s — it is comparing years of commission against the one-time and ongoing costs of the alternative, and every year that build cost falls, more of today’s customers may clear that bar. Booking and Expedia have both defended against this so far with the same tool — Genius and One Key, an Optional Component of the Price (Pattern: an add-on element, fees, bonuses, terms, or performance charges, that changes the net cash-equivalent price paid) aimed at the traveler side of the market rather than a cut to the headline commission itself — which is a smart way to protect revenue without a price war, but it does nothing to change a hotel’s build-versus-buy math. That is a different problem, and it needs a different answer.

That question gets sharper, not softer, if artificial intelligence turns out to lower the cost of building a direct-booking platform the way cloud infrastructure lowered the cost of building almost everything else this decade. And that is two different questions wearing one name. Whether AI-powered booking engines and automated payment handling shrink the engineering cost of building a direct channel is a supply-side question — it is about whether a mid-size independent can suddenly afford what only a Marriott could afford before. Whether AI travel agents change how a traveler finds a hotel in the first place is a demand-side question, and a different one. The two could resolve on very different timelines, and conflating them risks missing whichever one arrives first.

Google Already Tried This, and Quit

On the demand side specifically, there is a precedent worth taking seriously, because it is not hypothetical — Google already ran the experiment. “Book on Google” let a traveler finish a hotel reservation without ever leaving Google’s site. It launched in 2015 and Google shut it down in 2022, and its own stated reason is more interesting than “customers didn’t want it”: low usage from hotels and airlines and travelers as well. In Google’s own words, most people still preferred to book on a partner’s site, whether the hotel’s own or an OTA’s. When Google pivoted in 2021 to free referral links instead — sending the click to Booking or Expedia rather than keeping the transaction — it found that model drove more engagement than owning checkout ever had, and it has stuck with the advertising business since.

That is a company with more search leverage than any AI agent has today, choosing to get paid for the click rather than own the booking. It is real evidence, not just assertion, for the argument Booking’s own CEO makes against the AI threat: owning payments across 100-plus methods and 220 countries, and serving a supply base that is roughly 90% independent properties with no infrastructure to sell themselves, is a genuine moat. Google hit that moat directly and turned around. Nothing about a chatbot seems to have changed that math yet.

The Risk Google’s Precedent Doesn’t Cover

Here is where I want to be careful not to let a good answer to one question stand in for the answer to a different one. Booking’s moat argument, and Google’s own retreat, are both about the demand side: whether a traveler can be talked out of using a trusted booking site. Neither one says anything about the supply side — whether AI makes it cheaper for a hotel to stop needing that booking site in the first place. Those are separate questions, and only one of them has actually been tested.

I think this is a real risk, not a speculative one, and it deserves to be named as such rather than filed as a footnote. The whole reason only the largest chains have crossed the backward-integration tipping point described above is that building a booking engine, wiring up dozens of payment methods, and running it reliably has been expensive — Marriott-and-Hilton expensive, not independent-hotel-group expensive. AI-powered booking engines, off-the-shelf dynamic pricing, and automated multi-currency payment handling are exactly the kind of tooling that turns a multi-year engineering build into a service a mid-size operator can buy. If that cost keeps falling, the tipping point this update already flagged for the next tranche of customers arrives faster than Booking or Expedia’s current defenses — Genius, One Key, the payments moat — are built to handle, because none of those defenses touch a hotel’s build-versus-buy decision at all. A moat that stops a traveler from leaving does nothing to stop a hotel from leaving. I don’t think this is priced into either company’s commission strategy today, and it should be.

The Business Isn’t About Airlines Anymore

One thing this post got right from the start and still holds: this industry’s economic center of gravity moved from airline fees to hotel commissions years ago, and it is not moving back. Airlines made up a low single-digit share of Expedia’s revenue by 2021. The three-way fee fight that opened this post was, in retrospect, the last skirmish over a battlefield the industry was already abandoning.

Bottom Line

Business 2 today looks nothing like the fee war that opened this post, and that is the easy read. The harder, more useful read is that the market’s apparent stability is doing some real work to hide a live pricing problem: Booking and Expedia both earn strong returns while the industry’s largest customers are already building their way out of paying them, and the open question is not whether that continues but who is next. Two of this post’s three original protagonists no longer exist independently, absorbed because winning them back organically never penciled out. The industry’s most disruptive move of the last decade, Airbnb, came from outside Business 2 rather than a rival inside it. And the company best positioned to have ended Business 2 as we know it, Google, already tried, and chose to keep taking a cut instead.

That last fact is genuine comfort, but only against the threat it actually tested. The risk I’d flag going into the future isn’t an AI agent politely asking a traveler where to stay — Google already ran that experiment and Business 2 won it. It’s AI quietly cutting the cost of the thing only a Marriott could previously afford to build, arriving at the next tranche of hotel supplier faster than either company’s loyalty-program defenses are built to notice. Booking and Expedia have a real answer to the demand-side AI question. They do not yet have one to the supply-side version, and on the evidence in this update, that is the more dangerous of the two.

***