Booking's Moat Isn't a Network Effect. It's a $7.3 Billion Annual Toll It Has to Keep Paying.

In 2024, travelers booked more than 1.1 billion room nights through Booking Holdings, moving roughly $165.6 billion of gross bookings across its platforms.¹ That is a staggering river of commerce, and the company turned it into $23.7 billion of revenue and $5.9 billion of net income.¹ But before the first dollar of that profit was counted, the company spent $7.3 billion - about 31 cents of every revenue dollar - on marketing, just to make sure the travelers showed up at all.³ Hold that number. It is the whole argument.

The official story is that Booking is a classic two-sided network: more hotels attract more travelers, more travelers attract more hotels, and the wheel spins faster on its own. The corrected story is that the wheel does not spin on its own. It spins because Booking pays an enormous, recurring toll - to Google, to ad auctions, to brand spend - to keep the demand side full. The supply side is genuinely sticky. The demand side is rented.

A real network effect would make this number fall

Here is the test that separates a flywheel from a treadmill. In a true network effect, scale is its own reward: each new user makes the platform more valuable to the next, so the cost of acquiring that next user goes down. The flywheel runs on its own momentum. Booking's marketing line should be shrinking as a share of the business. It isn't. Marketing spend rose to $7.3 billion in 2024, up from $6.8 billion the year before⁸, and it still consumes about a third of revenue.³ That is not the signature of a self-reinforcing loop. It is the signature of a company that must re-buy its audience every quarter.

The reason is structural. Most people don't have a relationship with an online travel agency; they have a relationship with a search box. When a traveler types 'hotels in Lisbon' into Google instead of opening the Booking app, Booking has to win that click at auction - against Expedia, against the hotel's own site, against Google itself. Every traveler who arrives that way is a traveler Booking paid for. The company's own loyalty machinery, the Genius program, exists precisely to convert those rented visitors into direct, repeat bookers who cost nothing to reach the second time.² And it's working at the margin: full-year direct-channel room nights reached a mid-fifties percentage of the total in 2024, up year over year.³ But notice what that admits - nearly half of all room nights still arrive through paid channels. The moat is real. It is also metered.

The supply side is the part that's actually hard to copy

If the demand side is rented, the supply side is where the durable advantage lives - and it is formidable. Booking.com offers a sprawl of inventory no newcomer could assemble: its About page reports 28 million accommodation listings, a figure that includes individual rooms and units rather than distinct properties.⁶ The cleaner, SEC-filed number is that the company reached 7.9 million alternative-accommodation listings - homes and apartments - by the end of 2024.² That breadth is the genuine moat. A hotel in a Croatian fishing village or a Tokyo apartment knows that Booking sends it guests at a scale no rival can match, so it lists there first and prices there competitively. The supplier can't easily leave, because leaving means giving up the largest single funnel of foreign demand it has.

This is why the global picture matters more than the American one. In the United States, Expedia still leads online-travel-agency share at roughly 19.3%⁷ - Booking's structural dominance is concentrated in Europe and Asia, not its home market. Together, Booking and Expedia account for around 60% of all travel bookings across Europe and the U.S.⁷, and Booking.com led global travel web traffic with about 562.6 million monthly visits in 2024.⁷ The supply gravity is real. The question is whether the people who own the rulebook in Booking's strongest region will let it keep pulling.

Europe just declared the moat a public road

On May 13, 2024, the European Commission designated Booking.com as a gatekeeper under the Digital Markets Act - the seventh company on that list - and gave it six months to comply, with non-compliance carrying fines of up to 10% of global turnover, and 20% for repeat infringements.⁴ Booking confirmed the designation on its own compliance page.⁵ Read that against the moat anatomy and the threat is precise: the DMA targets exactly the levers that make Booking's supply gravity self-reinforcing - the price-parity pressure on hotels, the prominence of its own listings, the friction it can put between a hotel and its direct customers. The regulator is, in effect, ordering Booking to stop using its scale to deepen its scale. And it is doing so in the one region where Booking is strongest.

But $8.3 billion of EBITDA doesn't lie - so where's the catch?

The fair objection is that this is too gloomy. Booking generated about $8.3 billion of adjusted EBITDA in 2024, up 17%, on roughly $8 billion of operating cash flow²³ - hardly a business being eaten alive. And the bears have been wrong before: every supposed threat, from metasearch to Google's own travel ambitions, has so far broken against Booking's inventory wall. That's true, and it's the strongest case for the moat. A platform that processes $165.6 billion of bookings has real pricing power; merchant gross bookings even grew 27% year over year as the company shifted toward handling payments itself.² The honest counter, though, is that none of that contradicts the thesis - it confirms the half of the moat that holds. The supply side is paying off. It's the demand side and the rulebook that are being repriced. A moat that costs $7.3 billion a year to defend and is being narrowed by regulation in its best region is still a moat. It is just not the free, eternal flywheel the story sells.

Booking is one of the great businesses of the internet age, and the temptation is to file it under 'unstoppable network effect' and move on. The numbers tell a sharper story. The supply side is a wall built over nearly three decades that competitors cannot scale. The demand side is a tollbooth Booking pays at, not one it collects - $7.3 billion a year and counting. And in Europe, where the wall is highest, the regulator has started removing bricks on purpose. The genius of Booking was assembling the inventory. The vulnerability is that it never stopped having to buy the crowd - and now the people who set the rules want to make the crowd easier to take away.