← all stories
1996–2006acquired by Priceline for $133M · 2005

Booking.com

Bookings.nl → Bookings B.V. → Booking.com

An online hotel booking service that a recent university graduate from the Netherlands (by his own account in one interview, still a final-year student) launched with two servers under his desk in 1996, copying the idea from the American Hilton.com. In ten years it went from 18 friends as investors and 10 hotels to being bought by Priceline for $133 million in 2005, a deal since called one of the best in internet history: an agency commission with no prepayment and no risk for the hotel proved stronger than Expedia's merchant model, which the whole industry considered the correct one.

Founders Geert-Jan Bruinsma (founder of Bookings.nl, 1996) · Stef Norden (CEO of the merged company from ~2000) · Arthur Kosten (co-founder of Bookingsportal 2001, CMO from 2003) · Kees Koolen (operations director from the early 2000s, CEO 2008–2011)
Domains bookings.nl · bookings.org (historical) · booking.com
travelotamarketplaceagency-modelcommissionperformance-marketingseoaffiliatelocalizationpayments-anglebootstrap

In brief

Booking.com began as a side project of a recent graduate of a Dutch university (by another version from the same founder, a final-year student, see "Discrepancies") who in 1996 saw that the American Hilton.com already let you book a room online while in the Netherlands you could not, and decided to fix that with two servers under his desk and money from 18 friends. In nine years the company went from 10 hotels and a commission of five percent, set that way because he knew of no other figure, to the sale to Priceline for $133 million in 2005, a deal since called one of the best acquisitions in internet history: the "boring" model with no prepayment for the hotel beat the model of its competitors (Expedia), which the whole industry considered more correct.

How it started (the founders)

Bruinsma's student status at the time of the founding is a discrepancy that is not resolved even inside one and the same source (for the details and the correction of the status, see "Discrepancies"): the alumni page of the University of Twente says "1994: Bruinsma graduated... (TBK'94)," meaning that by 1996 he had been a graduate for two years; but in the same Skift interview that is quoted heavily in this dossier, Bruinsma himself describes the moment he found Hilton.com with the words "In my final year in university, I got in contact with the Internet...," that is, in his final year, still a student unverified. The origin of the company's idea is also told differently in different interviews. In the Skift version (Bruinsma's own words) he was searching the internet for a way to book a hotel, found no Dutch equivalent, and came across the American Hilton.com, in places borrowing design elements outright, and this "must have been July 1996"; in the version given by the newspaper Tubantia in 2016 the idea came at a dinner where a friend told him about a bad holiday, and the next day Bruinsma registered bookings.nl, with Hilton.com not mentioned at all in that account estimate. He had no experience in the hotel business whatsoever beyond a student job as a night porter—that was where he learned the mechanics of booking well enough to write the product. He raised the starting capital by emailing a business plan to a few dozen acquaintances who had internet access, mostly brothers from his student fraternity; 18 of them invested, each a modest amount. On November 12, 1996, he registered a legal entity under the name Boekingspunt Nederland: the chamber of commerce had rejected the application for the name "Bookings.nl." He set the commission for hotels at 5% with almost no deliberation: "I had no clue about commission rates; that's why I started with 5 percent." The same logic, that the hotel knows its own price better than he does, became a permanent principle of the company. By January 1997 the first 10 hotels had signed up, and the first booking followed soon after. Sources still name the place of the start differently, but the balance has shifted: summer/July 1996 now has three sources (UT Canon, the direct quote from Bruinsma in Skift, the Tubantia interview), while the city of Enschede has two direct independent sources (the university magazine U-Today and the 2016 Tubantia interview), both of which speak specifically about the headquarters moving FROM Enschede TO Amsterdam; UT Canon does not name the city, but does not contradict it either. Against that, September 1996 and Amersfoort appear only in Netkwesties, duplicated twice without independence (the source itself admits it is secondary). Summer/July 1996, Enschede is used as the better-confirmed version, with Amersfoort/September as the weaker alternative; the exact date of the headquarters move to Diemen/Amsterdam is still not established unverified.

In parallel and entirely independently, in September 1999 in Cambridge (UK) the cousins Andy Phillipps and Adrian Critchlow launched Activebooking.com, their own hotel booking site, copying part of the logic from the American WorldRes; in early 2001, after the dot-com crash, the company was renamed Active Hotels. This British company had nothing to do with the Dutch merger of Bookings.nl, and the two stories would come together only in 2006, by then under Priceline.

Year-by-year timeline

Lesser-known but significant facts

  1. The 5% commission was chosen not from a calculation but because nobody there knew any other figure. Bruinsma puts it plainly himself: "I had no clue about commission rates; that's why I started with 5 percent." That uninformed but modest start became the norm, and it held for years, until the market itself raised the rate.
  2. The affiliate model was born of weakness rather than strength. By 1997 Bookings.nl was only the THIRD largest Amsterdam hotel site, behind two email competitors that had no online booking at all; instead of competing, Bruinsma offered both of them a 50/50 commission deal for embedding his engine in their sites, and both agreed within the hour. The word "affiliate" was not even in his vocabulary then.
  3. The first marketing failure accidentally produced the SEO strategy. The newspaper De Telegraaf refused to publish an ad for Bookings.nl because it gave an internet address instead of a phone number; editorial policy in 1997 did not accept such a contact as legitimate. Bruinsma moved to paid search advertising on AltaVista (before Google), and even then 60% of traffic came from search.
  4. The British "subsidiary" physically handed out modems to get offline hotels onto the internet. In 1999 only ~40% of independent British hotels had computers, and only half of those had internet; Active Hotels distributed Alcatel "Web Phone" devices with a built-in modem to physically connect the first 500 partners—within a year the share of connected hotels rose from 20% to 80%.
  5. Booking.com's signature "urgency" notifications grew out of books about dating and social proof. In the early 2000s marketing director Arthur Kosten bought the product team books on consumer psychology and persuasion (in the spirit of pick-up techniques) in order to reverse-engineer trust signals such as the idea that people believe a crowded restaurant, and that is where the prototype of the "booked N times in the past day" mechanic came from.

The first growth lever

Growth rested on a bundle of three parts, each of them a forced decision rather than a strategic insight. The core was the agency commission model: the hotel sets the price itself, the guest does not pay in advance, and the commission is charged only AFTER the guest checks out, with no prepayment and no merchant risk for the partner. This was not a philosophical choice but a consequence of a lack of capital: at Britain's Active Hotels (the same model, arrived at independently) the choice between the agency and merchant schemes took up the company's first six months and was settled in favor of agency precisely because a startup had no money to become a merchant. More than that, when Active Hotels did test the merchant model in practice in 2002, the pilot failed on both sides: consumers did not want to pay by card in advance, and hotels disliked what prepayment did to their own cash flow. Active, profitable by then, went back to the agency model on the evidence rather than out of stubbornness. Priceline's own 10-K for 2005 formalizes this as the main structural difference: "not compensated by the hotel property until such customer checks out."

The second part of the bundle was distribution without building traffic of its own from scratch: an affiliate network from 1997–1998 (the Channels.nl entry above) and, in parallel, SEO landing pages for every destination and every hotel instead of one shared home page. When Google AdWords appeared in the early 2000s, the same infrastructure, which Arthur Kosten described as a factory process for producing hundreds of localized landing pages for a specific route (a language version → advertising in the language of the market → an A/B test → iteration), moved straight from organic to paid search. Booking.com became Google's largest customer in Europe, expanding so fast that "Google asked us if we could slow down a bit: our system was creating new campaigns faster than Google's servers could process them." Bruinsma himself, independently of Kosten, confirms the early (1997–98) practice: "We had landing pages before the term landing page existed... This was all before Google." Priceline's own form 10-K for 2005 confirms this dependence separately at the level of an official risk factor, now as a primary document from the buyer rather than a retelling of Kosten or Skift: "our European operations utilize online affiliate marketing and Internet search engines, principally through the purchase of travel-related keywords, as principal means of generating traffic," and, separately, the risk of price changes: "a significant amount of our European business is directed to our own websites through participation in pay-per-click advertising campaigns on Internet search engines whose pricing and operating dynamics can experience rapid change." A key detail: Bookings had a structurally LOWER margin than Expedia, and that was exactly what forced constant optimization of conversion instead of resting on its laurels.

The third part is the other side of the same coin: the agency model creates a cash gap. Google demands payment for advertising up front, while the commission from the hotel arrives only weeks or months after the guest checks out. By indirect evidence it was this structural gap—and not only a wish to exit gracefully—that made the deal with the well-capitalized Priceline happen at all: it "gave enough room to pay the bill to Google" estimate. The result by the time of the sale: ~$225M in gross bookings for the 12 months to mid-2005 (+110% year over year) and ~$25M in revenue with no profit; after the deal, in Q4 2005, the European segment grew to $158M in gross bookings with 88% organic growth, against just 5% for Priceline's own core US business in the same quarter.

Parallels today (projects from the catalog)

What a builder can take from this in 2026

  1. Not knowing the "right" number is no reason not to start. The 5% commission came from ignorance of the market rather than from a calculation, and it still worked, because it could be changed later as growth came in.
  2. A weak position in a local market is a reason to make a deal with competitors, not only to compete with them. As the third player in Amsterdam, Bruinsma got distribution by embedding himself in the sites of two larger rivals instead of trying to outrun them head-on.
  3. An industry consensus can be wrong for longer than you have patience, and that still does not make it right. In November 2003 the profession spent two hours explaining that the agency model was dead; the company grew by hundreds of percent a year in spite of that, and a decade later the CEO of Expedia admitted it himself: "we were attached to the merchant model... that blinded us."
  4. An alternative growth model is worth testing literally, not arguing about in a meeting room. Active Hotels did not philosophize about the merchant model; it launched it, saw the failure on real users, and came back to the agency model with evidence in hand.
  5. The mechanic that drives growth can be the company's structural risk at the same time. The same cash gap (pay the channel up front, collect revenue with a lag) that built Booking.com became one of the reasons the company had to be sold to a better-capitalized player at all.

Discrepancies and what we could not verify

Sources (primary first)

Primary (period documents, the founders' own words, official chronology):

Secondary (analysis and retrospectives, used with markers):

The same thing, about today

The same breakdowns, but of projects launching right now: what the product is, where the first users came from, how they charge. The card is free, the full dossier is $5 (the dossier itself is written in Russian).

Browse the dossier catalog →