The $135 Million Accident
In 2005, a Connecticut-based company called Priceline.com — whose stock had cratered 99% from its dot-com peak, whose original business model was a novelty act disguised as a revolution, and whose market capitalization had only recently crawled back above $3 billion — paid $135 million in cash for a small Dutch hotel-booking website called Booking.com. The deal barely registered on Wall Street. Priceline had been a punchline for half a decade, a poster child for late-1990s exuberance: William Shatner in television ads, a "Name Your Own Price" reverse-auction gimmick, a $20 billion valuation that evaporated overnight. The acquisition of Booking.com looked like a minor footnote in a minor company's attempt to find a second act in European hotel reservations. It was, by almost every measure that matters in retrospect, the greatest acquisition in the history of the internet.
Today, Booking Holdings Inc. (NASDAQ: BKNG) commands a market capitalization exceeding $170 billion. In fiscal year 2024, the company generated $23.7 billion in revenue and over $8 billion in free cash flow. Booking.com — that $135 million afterthought — constitutes roughly 90% of the parent company's revenue. It lists over 28 million accommodation properties in 43 languages. In 2023 alone, travelers booked more than one billion room nights through Booking Holdings' platforms, dwarfing Airbnb's 488 million and Expedia's 351 million. The company has quietly retired more than 21% of its outstanding shares over the past two years alone, compounding per-share value at a rate that would make even the most disciplined capital allocators envious.
The paradox at the center of this story: the greatest travel company in the world was built not by a Silicon Valley visionary or a hospitality mogul, but by a series of pragmatic, unsentimental decisions made by people who had already survived near-extinction and had no attachment to the narratives that nearly killed them.
By the Numbers
Booking Holdings at Scale
$23.7BFY2024 revenue
$172BMarket capitalization (mid-2025)
28M+Reported accommodation listings
1.05BRoom nights booked (2023)
~38%Free cash flow margin
$8B+Annual free cash flow
198Offices across 70+ countries
21%Shares retired in two years
The Penny Stock That Ate the World
To understand Booking.com, you first have to understand Priceline.com — specifically, you have to understand how wrong everyone was about it.
Jay S. Walker founded Priceline in 1998 around a patented mechanism he called "Name Your Own Price." The concept was elegant in theory: airlines and hotels carry perishable inventory — an empty seat or an unbooked room has a marginal cost approaching zero — and Priceline would let consumers bid blind for this surplus capacity. Suppliers could offload excess inventory at a discount without cannibalizing their transparent pricing channels. Consumers got a deal. Priceline got a cut. The friction was the point: you wouldn't know the airline or hotel until after you'd committed.
The numbers were absurd. Priceline reported $35 million in revenue in 1998, its first year. Revenue hit $482 million in 1999. By 2000, it had crossed $1.2 billion. The company went public in March 1999, and by May its market cap exceeded $20 billion. Then the dot-com bubble burst, and it became clear that revenue growth built on consumer novelty and venture-funded marketing is not a business model. By December 2000, Priceline's market cap had collapsed below $200 million — a 99% drawdown. The stock fell so far the company executed a 1-for-6 reverse stock split to avoid delisting. Nine months later, the September 11 attacks dealt a seemingly fatal blow to an already-gutted online travel company.
What surprised even the cynics: it took Priceline seven years to exceed its 2000 revenue. This was not a stock-price correction. The fundamentals cratered too, even as the internet proved itself as a genuinely revolutionary distribution channel. The lesson embedded in this history — one that would shape every subsequent decision the company made — was that being early to a real trend offers no protection if your specific business model is wrong.
The battle-hardened, pragmatic culture was perhaps the most valuable asset to survive the crash. It created a leadership team that was unsentimental about its original, failing business model and actively searched for one that actually worked.
A Student, a Gap, and a URL
While Priceline was burning through its second life in Connecticut, the real story was already underway in Amsterdam.
Geert-Jan Bruinsma was a Dutch university student who, in 1996, noticed something that seemed too obvious to be an insight: he could book a hotel room online in the United States through Hilton.com, but he couldn't do the same thing in his own city. Amsterdam — one of the most visited cities in Europe, a place where backpackers and business travelers and canal-tour tourists generated enormous and fragmented lodging demand — had no online hotel-booking platform. Bruinsma registered Bookings.nl and built a rudimentary website from a small Amsterdam office.
The timing mattered. In 1996, Europe's hotel market was vastly more fragmented than America's. The continent had hundreds of thousands of independent hotels, boutique properties, bed-and-breakfasts, and pensions that lacked the resources or technical sophistication to build their own online presence. The big American chains — Hilton, Marriott, Starwood — dominated the U.S. market with their own loyalty programs and direct booking capabilities, but in Europe, independent properties were the market. This fragmentation was simultaneously the problem and the opportunity: no single hotel could afford to build a global distribution platform, but any platform that aggregated enough of them could become the default channel through which travelers found accommodation.
Bookings.nl grew slowly. For almost a decade, it was a modest European startup, expanding city by city, hotel by hotel, with the unglamorous work of signing up individual properties, translating listings, and building the kind of localized, multi-language infrastructure that American competitors found tedious and unprofitable. The company eventually became Booking.com. By the time Priceline came calling in 2005, it had built something that looked small but was structurally formidable: deep supply in the world's most fragmented and highest-value hotel market, a merchant model that let hotels pay only for results, and a technology platform designed from the ground up for international complexity.
The Man Who Saw Amsterdam From Connecticut
Glenn Fogel was, in the mid-2000s, the person inside Priceline responsible for the company's small European operation. A Harvard Law graduate who had worked at CIBC and eventually joined Priceline's corporate development team, Fogel was the antithesis of a charismatic tech founder. He was analytical, cautious, and constitutionally allergic to the kind of narrative-driven optimism that had nearly destroyed the company. What Fogel saw in Europe was an uncomfortable truth: the "Name Your Own Price" model simply didn't work there. European hotels were too fragmented, too independent, too culturally diverse for a blind-auction mechanism designed for American chain hotels with standardized inventory. The model that did work — straightforward commission-based booking with transparent pricing — was exactly what Booking.com had built.
Fogel championed the acquisition of Booking.com. Priceline paid $135 million in cash. In the same period, Priceline had also acquired Active Hotels for $165 million and integrated it into Booking.com's platform, effectively doubling down on the European commission model. The combined price tag — roughly $300 million — was a rounding error compared to what the asset would become.
The critical decision was what happened next. Rather than absorbing Booking.com into Priceline's American infrastructure, imposing its brand, or forcing the "Name Your Own Price" model onto European consumers, Priceline's leadership — first under CEO Jeffery Boyd, later under Darren Huston, and ultimately under Fogel himself — let Booking.com run independently. The Amsterdam team kept its culture, its technology stack, its commercial approach. Priceline's contribution was capital and strategic air cover. Booking.com's contribution was everything else.
We had a very simple principle: don't screw up the thing that's working. The Booking.com team understood European travel in a way that nobody in Norwalk, Connecticut ever could.
The Agency Model and the Art of Frictionlessness
The engine that powered Booking.com's ascent was deceptively simple: the agency model.
In the traditional agency model, a traveler finds a hotel on Booking.com, makes a reservation, and pays nothing upfront. The traveler pays the hotel directly upon arrival. After checkout, the hotel sends Booking.com a commission — typically 15% to 25% of the room rate — roughly one month later. Booking.com bears no inventory risk. It holds no rooms. It employs no housekeepers. It is, in the purest sense, a marketplace that connects fragmented supply with global demand and takes a cut.
This model was perfectly calibrated for Europe's independent hotel market. For a 30-room boutique hotel in Bruges or a family-run pension in the Dolomites, Booking.com offered something previously unattainable: access to a global audience of travelers, in their own language, with their own payment preferences, without requiring the property owner to understand search engine optimization, digital advertising, or cross-border payment processing. The hotel paid nothing unless a guest actually showed up. The risk was entirely asymmetric — tilted in the property's favor, or so it seemed.
The catch, which would only become apparent over years, was that the model created profound dependency. Once a property received the majority of its bookings through Booking.com, the platform's commission became not a marketing expense but a structural cost of doing business — a toll levied on demand that the property could not generate on its own. The "best price" parity clauses that Booking.com enforced — requiring hotels not to offer lower prices on competing platforms or even their own websites — ensured that the platform's pricing advantage was self-perpetuating. Hotels that tried to undercut Booking.com on their own site risked losing their ranking on the platform, which controlled the lion's share of their demand.
The hotel industry noticed, eventually. By 2025, more than 10,000 European hotels had joined a class-action lawsuit brought by the Association of Hotels, Restaurants and Cafes in Europe (Hotrec), seeking damages for the period from 2004 to 2024. The lawsuit alleges that Booking.com's parity clauses were extracted under enormous competitive pressure and distorted the European hospitality market for two decades. Booking.com called the claims "incorrect and misleading," cited its own survey showing 74% of hoteliers say the platform made their businesses more profitable, and pointed to a 2024 European Court of Justice ruling that it argues did not find the clauses anti-competitive per se.
The tension is real and unresolved. The same platform that gave independent European hotels access to global demand also became the gatekeeper that extracted an ever-increasing share of their revenue.
The Culture of the Experiment
If the agency model was the economic engine, the cultural engine was something more unusual: a near-religious commitment to A/B testing and online experimentation at a scale that most technology companies still don't achieve.
Harvard Business School professor Stefan Thomke documented Booking.com's experimentation culture in a 2018 case study that became one of HBS's most widely taught cases. The picture he painted was striking: by the mid-2010s, Booking.com was running thousands of simultaneous experiments on its platform at any given time. Every change — the color of a button, the phrasing of a free-cancellation badge, the ordering of search results, the timing of a push notification — was tested against a control group with statistical rigor before deployment.
To unlock the potential of large-scale testing, the leadership team had to challenge conventional assumptions about culture, process, and the management of innovation.
The system was not merely a product-optimization tool. It was a management philosophy. Booking.com's leadership deliberately chose not to rely on executive intuition or past experience when designing digital experiences. Instead, they built an infrastructure that democratized decision-making: any engineer or product manager could propose and run an experiment without executive approval. The results — raw data on user behavior, conversion rates, and revenue impact — settled debates that in most organizations would be resolved by whoever had the most seniority or the loudest voice.
This culture produced two compounding advantages. First, it created a machine that continuously optimized conversion rates — the percentage of visitors who actually book a room — at a pace competitors couldn't match. Tiny improvements in conversion, compounded across billions of visits per year, translated into billions of dollars in incremental revenue. Second, it made Booking.com's institutional knowledge empirical rather than narrative. The company didn't have strong opinions about what worked; it had strong evidence about what worked. This made the organization unusually adaptable and difficult to outmaneuver, because competitors weren't fighting a strategy — they were fighting a learning rate.
The failure rate was staggering. Most experiments failed. Thomke's research suggested that the vast majority of tested ideas produced no measurable improvement or made things worse. The culture required comfort with — even celebration of — failure. An engineer whose experiment showed a negative result had still contributed valuable information. The organizational design rewarded the process of rigorous testing, not the outcome of any individual test.
By 2011, just six years after the acquisition, Booking.com was generating more than $1 billion in annual profits — making it, according to HBS, the most financially successful digital travel marketplace in the world.
The Merchant Pivot and the Connected Trip
The agency model built the empire. The merchant model is building the next one.
For most of Booking.com's history, the platform did not process payments. The traveler booked, the traveler paid the hotel directly, the hotel paid Booking.com a commission after the fact. This was clean and capital-light, but it left money on the table — literally. When Booking.com didn't touch the payment, it couldn't earn float on customer funds, couldn't easily upsell additional services at the point of purchase, and couldn't build the kind of end-to-end transaction data that powers recommendation engines and dynamic pricing.
The shift to a merchant model — in which Booking.com collects payment from the traveler upfront, holds the funds, and remits to the property after deducting its commission — has been one of the most important strategic transitions of the past decade. By FY2024, merchant revenues constituted approximately 59.6% of total revenue (~$14.1 billion), up from roughly 50% just a few years earlier. Agency revenues, the original model, had declined to about 35.9% (~$8.5 billion). The remaining ~4.5% (~$1.1 billion) came from advertising and other services, the fastest-growing segment at 8%+ year-over-year.
The merchant model does several things simultaneously. It gives Booking.com control of the payment flow, enabling the company to earn interest on customer funds held between booking and checkout. It creates a natural upsell environment — once a traveler has entered payment credentials, adding a rental car, a flight, an airport transfer, or a museum ticket is a single click away. And it generates richer transaction data that feeds Booking.com's AI and personalization systems.
This is the architecture behind what the company calls the "Connected Trip" — its strategic vision of becoming not just a hotel-booking platform but a comprehensive travel marketplace. Booking.com now facilitates airline ticket bookings (up 33.1% year-over-year in Q1 2024), rental car reservations (up 10.7% in the same period), attractions, insurance, and ground transportation. The goal, articulated by CEO Glenn Fogel with characteristic bluntness, is to recreate the experience of having a human travel agent — someone who knows your preferences, handles the logistics, and fixes problems when things go wrong — through AI-powered software.
✈️
The Connected Trip Vision
Booking.com's evolution from hotel booking to full-trip marketplace
1996Bookings.nl founded in Amsterdam — hotel reservations only.
2005Acquired by Priceline for $135M. Agency model dominates.
2011Exceeds $1B in annual profits. Experimentation culture at scale.
~2018Merchant model transition begins. Payment processing integrated.
2023Launches AI Trip Planner using OpenAI's ChatGPT API.
2024Merchant revenue surpasses 59% of total. Flights, cars, attractions live.
2025Smart Filter launched. Internal AI orchestration layer spans OpenAI, Anthropic, Google.
Glenn Fogel and the Anti-Visionary
Glenn Fogel became CEO of Priceline Group (as it was still known) in 2017, after Darren Huston resigned following a violation of the company's code of conduct related to a personal relationship. Fogel's ascent was not the product of a founder's charisma or a board's visionary bet. It was, like much of the company's history, pragmatic — the right person in the right seat at a moment that demanded operational discipline rather than dramatic reinvention.
Fogel's biography reads like a deliberate counternarrative to Silicon Valley mythology. He survived a stroke at age 17. He trained as a lawyer. He worked in banking. He joined Priceline in its darkest years and spent more than a decade in corporate development, executing deals — including the Booking.com acquisition — that were tactical rather than transformational in their framing but world-altering in their compounding effects. At a 2024 interview with the BBC, Fogel described experiencing the dot-com crash, 9/11, SARS, and the 2008 financial crisis as the formative events that shaped his leadership: a bone-deep skepticism of hype, a preference for evidence over intuition, and an almost obsessive focus on the mechanics of capital allocation and shareholder value.
Under Fogel, the company rebranded from Priceline Group to Booking Holdings in 2018 — a belated acknowledgment that the tail had been wagging the dog for over a decade. He orchestrated the shift to the merchant model, championed the Connected Trip strategy, and maintained the company's extraordinary discipline around share repurchases. Booking Holdings has retired an astonishing number of shares: the company's float has been shrinking year after year, concentrating value for remaining shareholders with a relentlessness that recalls the Singleton-era capital allocation playbooks.
He insists on leaving for the airport hours early. He has never used a travel agent. He presides over a company valued at $172 billion with the disposition of someone managing a well-run utility.
The Airbnb Question
The most persistent narrative challenge Booking Holdings faces is the assumption that Airbnb is disrupting it.
The numbers tell a different story. In 2023, Booking Holdings booked 1.05 billion room nights. Airbnb booked 488 million. Expedia booked 351 million. Booking is not merely larger — it is growing faster in absolute terms. Alternative accommodations (homes, apartments, and other non-hotel listings) represented approximately 33% of Booking.com's room nights in 2023, up from 30% in 2022, and the platform now lists over 6.6 million homes, apartments, and unique places to stay alongside its traditional hotel inventory. Booking.com is competing directly with Airbnb in alternative accommodations while simultaneously maintaining its dominance in the hotel category where Airbnb has minimal presence.
The structural advantage is supply breadth. A traveler searching on Booking.com sees hotels, hostels, apartments, villas, and boutique properties in a single search — a comprehensive view of available accommodation that Airbnb, by design, cannot replicate (it has no significant hotel inventory). Expedia can offer both, but with less supply depth in international markets, particularly Europe and Asia. Booking.com's position as the platform that shows everything in a single search creates a natural gravitational pull for the undecided traveler — the person who doesn't yet know whether they want a hotel or an apartment, who is comparison-shopping across formats, and who will default to the platform with the widest selection and the best conversion experience.
One of my jobs is to make sure that we make ourselves as aware to the U.S. traveler who wants to find a home in the U.S. as any other part of the world. Our room-night growth rate in the U.S. is something I'm very focused on.
The real competitive anxiety for Booking.com is not Airbnb but Google. Google controls the search funnel through which many travelers begin their journey. It has been steadily building its own travel products — Google Hotels, Google Flights — and directing an increasing share of travel queries toward its own tools before users ever reach an online travel agency. The EU's Digital Markets Act designation of Booking.com as a "gatekeeper" in May 2024 (the European Commission classified its hotel reservation service as a "core platform service" under the DMA) reflects the company's market power but also highlights the regulatory pressure that comes with it.
The AI Wager
The most consequential bet Booking.com is making in 2025 is on generative AI — not as a marketing gimmick but as the architectural foundation for the Connected Trip experience.
CTO Rob Francis, who joined in 2019 after stints at Amazon, Sonos, and JPMorgan Chase, has been the internal champion of a measured but aggressive AI deployment. Francis was wary of the generative AI FOMO that swept the technology industry in late 2022 and early 2023 — the rush to bolt ChatGPT plugins onto existing products without a clear user experience thesis. Booking.com took a deliberate pause. "We didn't think that was a particularly great experience," Francis told Fortune in April 2025. "And so we took a little bit more time to think of what a richer experience might be for our customers."
The results of that pause have been substantive. In 2023, Booking.com launched its AI Trip Planner, leveraging large language models — including OpenAI's ChatGPT API — to answer freeform travel questions. By 2025, the company had launched Smart Filter, which allows travelers to type natural-language queries ("Hotel in Paris, with a rooftop bar and an indoor pool") and have an LLM scan Booking.com's inventory to surface matching properties. Internally, Booking.com built its own orchestration layer that allows product and engineering teams to switch between OpenAI, Anthropic, Google, and open-source models without re-engineering their applications — a bet on model optionality in a rapidly evolving market.
The company's own research found that 48% of travelers would trust AI to plan a trip for them. Francis says Booking.com monitors how many room nights are booked using AI-powered tools but has deliberately avoided setting firm targets — a posture consistent with the experimentation culture. "We're still learning the best way to have the right outcome for an experience," he said.
The deeper strategic logic: if generative AI truly can replicate the human travel agent experience — understanding preferences, suggesting itineraries, handling changes and cancellations through conversational interfaces — then the platform with the deepest supply, the richest transaction data, and the most sophisticated AI capabilities will capture a disproportionate share of travel intent. Booking.com is betting that this platform is itself. The risk, articulated sharply in a January 2026 Harvard Business Review article, is that generative AI could instead disintermediate platforms like Booking.com entirely — allowing new AI-native entrants to aggregate supply directly and bypass the established OTAs.
The Gatekeeper and Its Gatekeepers
Booking.com's very success has made it a target.
In May 2024, the European Commission officially designated Booking Holdings as a "gatekeeper" under the Digital Markets Act, classifying Booking.com's hotel reservation platform as a "core platform service." The designation places Booking.com alongside Apple, Google, Meta, and a handful of other dominant platforms subject to heightened regulatory scrutiny, including a list of do's and don'ts — such as prohibitions on self-preferencing and requirements to give business users fair access — enforced by the threat of fines up to 10% of global revenue.
European Commission Executive Vice President Margrethe Vestager said the decision meant that "vacationers will start benefiting from more choice and hotels will have more business opportunities." Booking.com said it had "been working with the European Commission for some time as we anticipated today's decision" and committed to compliance within the six-month window the regulation requires.
The DMA designation was, in a sense, the regulatory codification of what the hotel industry had been arguing for years. The parity clauses that locked hotels into Booking.com's pricing framework were dismantled — Booking.com dropped them in 2024 to comply with the DMA — but the class-action lawsuit filed by Hotrec seeks damages for the two decades before the clauses were eliminated. The hotel industry's argument is that the damage was structural: Booking.com used its early market power to prevent the emergence of competitive alternatives, and the market concentration that exists today is the lasting product of those practices.
Booking.com's defense — that parity clauses fostered competitive pricing rather than restricting it, and that the vast majority of hoteliers found the platform profitable — is plausible but not unassailable. The lawsuit, expected to be one of the largest ever filed in the European hospitality sector, will be heard in Amsterdam, the city where Geert-Jan Bruinsma once registered a URL because he couldn't book a hotel room online.
The Capital Allocation Machine
If there is a single number that explains why Booking Holdings trades at $172 billion despite generating the kind of margins and growth typically associated with much smaller companies, it is the share count.
Booking Holdings is, behind the travel marketplace, one of the most aggressive share repurchase programs in public markets. The company has been buying back its own stock for over a decade, and the pace has accelerated. Over the last two years alone, the company retired approximately 21% of its outstanding shares. Combined with the introduction of a dividend — a relatively recent addition — the result is a capital return program that channels nearly all of the company's $8 billion+ in annual free cash flow back to shareholders.
The math is straightforward but worth spelling out. Revenue grows at a low-to-mid-teens rate. EBITDA margins have expanded sequentially in every single quarter for the past three years.
Free cash flow margins hover around 38%. The company has minimal need for reinvestment capital — it owns no hotels, operates no airlines, and holds no physical inventory. The vast majority of capital expenditure goes to technology infrastructure and marketing. What's left over — and it is an enormous amount — goes back to shareholders, primarily through buybacks that reduce the denominator of every per-share metric.
The effect is compounding at an extraordinary rate. Earnings per share grow faster than net income. Free cash flow per share grows faster than free cash flow. The stock price, over any reasonable time horizon, has reflected this: Booking Holdings has been one of the best-performing large-cap stocks of the past two decades, a fact largely invisible to the public because the company doesn't court media attention, its CEO doesn't post on social media, and its product — unlike an iPhone or a Tesla — is something people use without thinking about the company behind it.
In April 2020, at the nadir of the COVID-19 pandemic — when global travel had effectively ceased — Booking Holdings issued $3.25 billion in senior notes at rates between 4.1% and 4.625%, raising a war chest that simultaneously fortified its balance sheet and signaled to investors that the company expected to survive what was, for the travel industry, a near-extinction event. The company's revenue dropped precipitously and didn't fully recover until 2022. But the balance sheet held. The experimentation culture continued. The share buybacks resumed. And by 2023, Booking Holdings was posting record revenue of $21.4 billion, 25% above the prior year — a recovery powered not by pent-up demand alone but by the structural advantages the company had quietly deepened during the crisis.
The Machine
The essential story of Booking.com — the one that explains the $135 million acquisition that became worth more than $150 billion, the experimentation culture that generates thousands of simultaneous tests, the agency model that locked in European supply, the merchant pivot that unlocked the Connected Trip, the capital allocation discipline that compounds per-share value — is a story about institutional learning.
The company's history, documented compellingly in Lukas Vermeer and Dan Siroker's
The Machine, reveals an organization that optimized not for any single strategic insight but for the
rate at which it generates and acts on insights. The experimentation infrastructure is the machine. The A/B tests are the inputs. The conversion improvements are the outputs. And the compounding effect of thousands of small improvements, applied across billions of sessions per year, creates an advantage that no competitor can replicate by copying any individual feature, because the advantage is not in any feature — it is in the velocity of learning itself.
Stefan Thomke's HBS case study reached a similar conclusion: Booking.com's competitive advantage was not its technology (replicable), not its supply (accessible to well-funded competitors), not its brand (Expedia and Airbnb are equally well known), but its culture of empirical decision-making — the institutional refusal to accept opinion as evidence, hierarchy as authority, or past success as a predictor of future performance. The scientific method, applied not in a laboratory but across a marketplace handling billions of dollars in transactions, with the discipline to let the data override the CEO.
There is something almost anti-literary about this kind of advantage. It lacks the drama of a charismatic founder, the elegance of a single strategic insight, the romance of a near-death pivot. It is, instead, the patient accumulation of fractional improvements — a thousand engineers running experiments on button colors and cancellation policies and push-notification timing — that compounds into something no competitor can see across the table and understand how to beat.
Fogel, asked at a 2024 Fortune interview about Booking Holdings' competitive position, described the future he's building toward: recreating the human travel agent experience through technology. "People who have had the pleasure of using a human travel agent actually remember that it was a really good experience," he said. "We want that back, only we want it better. Because, yeah, the travel agent knew you kind of, but the database with all the previous things you've done knows you much, much better."
The 51% of room nights now booked on the Booking.com mobile app — up from 46% just one year earlier — is the quiet metric that carries the weight. App bookings are direct. They bypass search engines. They represent a customer who has chosen Booking.com as their default interface with the world of travel. Every percentage point of app share gained is a percentage point of Google dependency shed, a percentage point of customer acquisition cost eliminated, a percentage point of the flywheel spinning faster.
In Amsterdam, in the city where a university student once noticed he couldn't book a hotel room online, 198 offices in 70 countries hum with the daily work of running experiments, signing up properties, processing payments, and training AI models — the infrastructure of a company that has compounded a $135 million acquisition into one of the most valuable internet businesses on earth, one fractional improvement at a time.
