Expedia ran 1,750 A/B tests last year and used two-way mirrors to study users' eyeballs — literally — and behavior as they meandering around site pages. It is a formula that turns into conversion — i.e. travelers clicking the book button.
On a Monday morning in mid-October, in a suburb of Seattle, a young woman named Megan went online to make some travel plans.
She and her parents, along with her siblings and their spouses, wanted to go somewhere tropical in January, and in a flurry of texts and Facebook messages, Belize had emerged as the leading candidate. It had fallen to Megan, as it often did, to execute. So, a little after 9 a.m., she typed the name of the online travel agency Expedia into her browser bar and began to explore flights.
Her preference was for Alaska Airlines — she’d had good luck with the carrier — but when she couldn’t find anything, she started looking at American. She noticed there was only one ticket left for the least expensive flight, which caused her some concern. She grew more apprehensive as she noticed that the cheaper available flights had long layovers. Then she saw that some of the layovers were in Los Angeles, and she briefly considered a visit to Disneyland.
After eight minutes, without settling on a flight, Megan began to explore hotels. The photos from a “jungle spa” resort caught her eye — it looked adventurous but also pampering — so she was crestfallen when she noticed that it was booked up for the dates she wanted. “Oh, so sad,” she said softly. She looked through the reviews of another promising hotel and found it had no Wi-Fi, which wouldn’t be a problem, but also ants, which would. She stumbled onto a review someone had written about going to a resort to recover after “a surprising end to a marriage engagement.” That bummed Megan out a bit.
Then she found what looked like the one. This hotel wasn’t on the beach, but the reviews mentioned a spiral staircase, which sounded neat, and an on-site bakery. Megan loved bakeries. And it wasn’t too expensive. She’d have to confirm with her family, she reminded herself, but it was pretty close to ideal.
At that moment, a disembodied woman’s voice came over a speaker and told Megan she was finished. The voice belonged to an Expedia user-experience researcher named Susan Motte, who, with a team of programmers and designers from the company’s hotel-shopping and activities-booking teams, had been sitting in the next room watching Megan through a two-way mirror. Megan, who actually was planning a family trip to Belize (and whose full name Expedia asked not to be used, for privacy reasons), had been invited to the Usability Lab at Expedia headquarters in Bellevue, Wash., compensated with a gift card, and asked to use the site as she would at home.
An eye tracker mounted on the bottom of the computer monitor logged where Megan was looking on the screen at any given moment. Sensors on one side of her face measured the electrical impulses in two muscles—the zygomaticus major, which tugs the corner of the mouth into a smile, and the corrugator supercilii, which furrows the brow. Megan’s emotions, manifested in infinitesimal changes in muscle fiber tone, had been playing out on a screen mounted on the wall in the adjoining room. A red waveform on a scrolling graph tracked her tension during the session, a green waveform below it, her delight.
All of her reactions, and her answers to the questions Motte asked as Megan used the site, went into a growing database. Expedia, the parent company of more than a dozen travel-oriented brands in addition to Expedia.com, is obsessed with figuring out how to make booking travel online more intuitive, more efficient, and more enjoyable. That means, among other things, understanding the psychodrama of trip planning: the shifting desires and paralyzing wealth of choices, the unsettling gyrations in room rates and ticket prices, the competing demands of family members and budgets and schedules, the need to balance the thirst for adventure against the fear of Zika virus in Latin America or Islamic State in Europe.
There’s a modest body of literature on the psychology of vacations, and one of its findings is that much of the pleasure comes from anticipation—a 1997 study found that people are happier thinking about a trip beforehand than when they’re actually taking it. The goal of Expedia’s usability researchers is not only to make Expedia’s various sites and mobile apps more efficient but also to make them an extension of the vacation fantasies that are always running in the back of our heads.
The basic act Megan performed under such close scrutiny plays out tens of millions of times a day—in homes, in offices, in line at the coffee shop on smartphones. In 2015 people performed 7.5 billion airfare searches and booked 203 million hotel room nights through Expedia and the other sites owned by the company—Hotels.com, the “meta-search” site Trivago, the business travel site Egencia, the discount site Hotwire, Australia’s Wotif, and others.
Over the past decade revenue has more than tripled, from $2.1 billion in 2005 to $6.7 billion last year, and the stock price has risen fivefold. Along with Priceline, whose sites include Priceline.com, Kayak, and OpenTable, Expedia dominates the online travel business.
Like Priceline, Expedia is a survivor from an earlier tech era — both date from the mid-1990s — but unlike contemporaries such as Netscape and Yahoo!, they continue to thrive. Both were fortunate enough to get big quickly in a business that benefits greatly from scale and brand recognition, and their growth over the past several years has correlated with a long, gradual economic expansion during which both vacation and business travel have steadily grown. Both have made big acquisitions — most of Priceline’s growth has come from Booking.com, the dominant European hotel website, which it purchased a decade ago. Expedia, for its part, last year acquired the original online travel company, Travelocity, for $280 million, the upstart Orbitz for $1.3 billion, and the vacation rental site HomeAway for $3.9 billion.
What distinguishes Expedia is its dedication to understanding the psyche of the modern travel planner. That may be most apparent in the Usability Lab, but much of it happens on the sites themselves, as the company relentlessly tests new ideas about look and feel and function. Expedia’s leadership believes the resulting base of knowledge is what has allowed it to recover from earlier stumbles. It’s also what they’re banking on to keep the company growing in a more hostile climate, as investors worry about the threat of Airbnb, consolidation in the hotel industry, and the ever-present specter of Google moving into the travel business. Then there’s the question of whether people are going to feel like traveling at all in what looks like a very tumultuous world.
Expedia wasn’t originally an online travel agency. In 1994, Rich Barton, a young Microsoft engineer, was put in charge of a project to create an encyclopedic travel guide on the then state-of-the-art technology of CD-ROM. “I pretty quickly decided that was a dumb idea,” he says. Barton persuaded Bill Gates to instead let him create a website where people could book trips rather than just research them.
The travel technology company Sabre, which was owned at the time by American Airlines and handled reservations for many of its competitors, was working on Travelocity, and got to market first. But Travelocity, hamstrung by its airline ownership, was slow to move from flights into the higher-margin business of hotel rooms. In 1999, at the height of the tech bubble, Expedia was spun off, with Barton as CEO. Microsoft retained a majority ownership stake.
“It was really one of the first businesses to use the Internet as a service for consumers,” recalls Barry Diller, now Expedia’s chairman and senior executive. “I couldn’t wait to buy it.”
Diller had already run Paramount Pictures and Fox in his previous career as an entertainment mogul. In 1999 he was in the process, through a series of sales and acquisitions, of turning a collection of TV stations into an Internet company. USA Interactive, as it was called at the time, purchased Microsoft’s stake in Expedia in 2002 for around $1.3 billion, then bought the rest of the company the next year. In 2005, Diller’s company, now called IAC/InterActiveCorp, spun off its profitable travel brands — Expedia, Hotels.com, TripAdvisor, and Hotwire — under the umbrella of Expedia Inc., with Diller retaining control of a majority of the shareholder votes.
The new Expedia struggled. Hotel and airline partners resented the company for the rates it demanded and started to take steps to circumvent it. Angered that consumers could often find lower prices on Expedia than on their own sites, some hotels ended their relationship with it. The major U.S. airlines started up a rival, Orbitz.
Most damaging, though, was the rapid rise of Booking.com and its competing business model. On Expedia, customers had to pay for hotel rooms when they booked. This so-called merchant model relied on buying large blocks of rooms at wholesale rates and marking them up. The margins were large; for the privilege of getting their wares into the giant global shop window that was Expedia, hotels often had to pay 25 percent or more of the room rate. Booking.com, on the other hand, used what was called the agency model: It allowed users to pay when they checked out (and it took a smaller cut in the form of a commission). When customers got a taste of that flexibility, they tended to prefer it. With customers leaving, revenue stagnant, and commissions squeezed, Expedia in 2005 faced obsolescence.
The goal is to make Expedia’s sites and apps an extension of the vacation fantasies that are always running in the back of our heads.
The company’s CEO, Dara Khosrowshahi, 34 at the time, was painfully aware that he didn’t know exactly how the company should respond to its new troubles. The Iranian-born former investment banker, who’d been IAC’s chief financial officer, decided he didn’t have to. If Expedia could be remade as a different sort of company, its customers would provide the answers themselves. At the time, Expedia was centrally organized, in terms of both people and software. Technology decisions for all sites ran through one chief technology officer; marketing questions ran through one chief marketing officer. Each site was an interconnected tangle of code. If programmers wanted to add the ability to handle credit card billing in Germany, for example, that might end up causing errors in the flight-search feature. As a result, changes were rare: A site would be updated only a couple times a year. Each change “would break the system down between a week and God knows what,” Diller says in an interview at IAC’s Manhattan headquarters. “We couldn’t change a Tuesday to a Thursday without thousands of hours of work.”
Khosrowshahi made the case to Diller that the company needed to spend hundreds of millions of dollars hiring and equipping coders to do work that would be largely invisible to consumers. “We just had to go in and invest very aggressively without having a clear return on investment,” Khosrowshahi recalls. Diller bought it, and he persuaded the board to buy it, too. The company’s tech spending increased from $130 million in 2005 to $362 million in 2010 — in 2015 it was $750 million. Brand Expedia alone has increased its coding workforce since 2010 from 200 software engineers to 2,000.
Today each of Expedia’s brands has its own technology and marketing teams, and they’re encouraged to set their own course. They all benefit from the massive inventory of hotel rooms and plane tickets and the financial resources and technological firepower of the parent company. How they use all that is up to them. “They are competing against each other, they steal from each other,” says Khosrowshahi. We’re sitting in a meeting room near his desk at Expedia headquarters in Bellevue, next to a whiteboard that he gradually fills with drawings and charts. “Sometimes they do the same thing as each other, which is kind of a waste,” the CEO concedes. “You know, they’ll build the same exact feature twice. But the trade-off is speed.”
The pace at which the sites evolve is paramount to him, and the most important measure of it at Expedia is how many tests the software engineers run. The company’s gospel is “test-and- learn,” pronounced as a single three-syllable word. Also called A/B testing, it borrows the logic of pharmaceutical research: Test an innovation against the status quo and see if there’s any difference. It’s common practice at Google and Facebook, but it was relatively unknown in the online travel industry until Booking.com started using it. Expedia’s commitment to the practice is quasi-cultish. In 2011, when the company’s chief product officer, John Kim, started there, Expedia was running 50 to 100 tests a year. Last year it ran 1,750.
Today Expedians, as they call themselves, don’t debate questions, they test them: How big should a button on the website be? Should hotel availability be presented in list form or calendar form? How likely is it that someone typing a particular travel-related search query into Google is booking a trip right then, and how much should Expedia therefore pay for an ad that will run in the search results page? What should the Expedia brand’s loyalty program look like, and should Hotwire have one at all? Are people more likely to reserve a hotel room if they’re shown how few are available at a particular price? (As it turns out, yes, even though they say they hate it.)
Expedia’s decision, for example, to preserve the distinct brands it purchases rather than subsume them into the Expedia.com site was driven by research showing that even in the world of online comparison shopping, people retain stubborn brand loyalty. According to Aman Bhutani, president of the Expedia brand, the same person might shop for a cheap business hotel on Travelocity, a nicer hotel for a family vacation on Expedia.com, and a car rental on Hotwire, even though all three are Expedia sites offering identical inventory. In keeping with the radical agnosticism of the company, Bhutani isn’t particularly interested in why that is. “I just want to know what you like and how many other people are like you so that I can provide those services,” he says.
Two-thirds of the A/B tests Expedia runs show no effect or a negative effect, and most of the successful ones are only marginally so. “It’s a game of inches,” Khosrowshahi says. “What you have now is lots and lots of inches, and it adds up to very big numbers.” This stockpile of technological and behavioral information allows the company to give its brands a far more nuanced understanding of what customers want and how to get them from window-shopping to booking. “The entire online travel market has improved its conversion rates,” says Ryan Williams, an analyst at Millward Brown Digital, a consulting firm that tracks Internet use. “But Expedia’s conversion improvement is outpacing the market.”
Expedia sees its traveler-behavior data as a defense against superfluity. Despite its size and name recognition, Expedia doesn’t have its own hotel rooms or its own airplanes. It’s a middleman among other middlemen. Booking.com still has the bigger inventory of hotel listings; Google shows flight and hotel results right in the search page. Even TripAdvisor, spun out of Expedia in 2011, has become a major competitor by allowing visitors to book directly on its site. For Expedia to continue to justify its commissions and its existence, it has to be better than everyone else at divining vacationers’ desires. That draws on both the steady churn of testing and learning and the more open-ended qualitative work in the Usability Lab, where researchers use terms like “delight” and “fantasy” in a clinical manner.
“We’re doing more work to understand the dream state, the plan state,” says Scott Jones, who runs Expedia’s design and user-experience department. “People are always planning a trip. They may not be actively shopping for a trip, but they’ve always got this idea in mind.” The problem, as he sees it, is that purchasing the tickets and reserving the room breaks the spell. “Booking travel online has become so full of decisions and answering questions,” Jones says. “People are always daydreaming about trips, but as soon as they come to our site, it becomes more of a work thing.”
The problem is, purchasing the tickets and reserving the room breaks the spell
Every December the annual Expedia Partner Conference takes place in Las Vegas, the same weekend as the National Finals Rodeo. Cruise-line execs and car rental company reps mix along the Strip with honest-to-God cowboys and cowgirls. The conference is a chance for people who do business with Expedia to meet with one another and with the company’s leadership. It’s also a time for Expedia to convince all of those partners that the company’s no longer the bully of a decade ago. Many of the new features Expedia is rolling out are for its suppliers, allowing them, for example, to monitor what sort of photos or descriptive language attracts the most Expedia users. For the past few years, Expedia has reduced commissions, sacrificing profits to increase its inventory by luring more hotels to work with it.
One of the conference’s main events was a conversation between Khosrowshahi and Diller. At the front of a cavernous ballroom at the Bellagio Hotel & Casino, the two sat onstage in leather club chairs. Khosrowshahi wore tight jeans and a slim, short suit jacket; Diller wore a pinstriped suit, driving shoes, and no socks.
The talk ranged from the future of cable television to the health effects of Coca-Cola to Donald Trump — “an evil man,” Diller called him. “I’m very proud that for 30 years I haven’t spoken to him, because I’ve always disliked him.” Midway through they turned to the question of whether people were too scared to travel. The preceding weeks had seen both the Paris and San Bernardino terror attacks. Islamic State was promising more. Trump was telling sellout crowds he wanted to deport 11 million illegal immigrants. In Europe, citizens were revolting against the idea of welcoming waves of Syrian refugees into their countries, and in the U.S., politicians and presidential candidates were rejecting the idea of letting in any at all. The world was feeling hunkered-down.
Khosrowshahi asked Diller what it meant for the travel industry, and the two started talking about Sept. 11. The attacks, along with everything else they wrought, caused people to stop traveling almost entirely for months. Diller was closing the purchase of Expedia at the time, and he and Khosrowshahi had debated whether to back out.
“We sat around and talked,” Diller said into his microphone, “and I don’t remember, you may remember who it was who said, ‘If there’s life, there’s travel.’ ”
“I think that was you,” Khosrowshahi interrupted.
“Oh,” Diller said. “So nice to hear. I’ve always given credit to someone else.”
It’s a resonant line, and over the day and a half that remained of the conference, various Expedia executives worked it into their presentations. It’s comforting to the company to think that the current trepidation about the world beyond one’s borders will pass. It’s comforting, too, to think of travel as a fundamental human activity, no matter what happens. Even among civilization’s radioactive remains, Mad Max will need to get from one place to another. He might need to find somewhere to stay when he arrives. And if there’s an on-site bakery, who’s to say he won’t want to know about it?
This article was written by Drake Bennett from Bloomberg and was legally licensed through the NewsCred publisher network.
Photo credit: Expedia's website is more science than art. Expedia