Steve Hafner and the unfinished business of going somewhere

In 2016, Steve Hafner thought he might have found his last job. He had helped launch Orbitz, cofounded KAYAK and stayed to run it after the sale to Priceline. A third startup held little obvious appeal. He was content to keep working on the travel search company he already had, preferably finishing on a high note. It was a reasonable plan. It also left quite a lot of room for the future to misbehave.

Ten years later, the plan has a sequel. In February 2026, Hafner handed KAYAK's chief executive job to Peer Bueller and became executive chair. By September, he and KAYAK cofounder Paul English were introducing Lola, an AI assistant backed by Booking Holdings. The two founders had another product to build together, and another set of assumptions to question.

The question had become smaller in geography and larger in ambition. KAYAK helped people compare ways to get somewhere. Lola starts earlier, with the unresolved weekend, the dinner reservation, the vague desire to go out. A person can know they want a good evening without having settled on a restaurant, a neighborhood or even a plan. That indecision is now part of Hafner's working territory.

His career has a pleasingly inconvenient continuity. Every time the internet makes travel look solved, another awkward piece remains. Comparing flights was one piece. Making a recommendation turn into something a person can actually reserve is another. Hafner has spent much of his professional life close enough to those gaps to find them irritating.

Ten countries before twelve

He encountered travel long before he encountered the travel business. Born in Lima to a Swedish mother and an American father from Arkansas, Hafner lived in ten countries during his first twelve years. Costa Rica, Guatemala, Sweden and several American cities figured in that childhood. Moving was a family condition before it became a commercial opportunity.

He has described the experience as an education in alternative ways of doing things. A familiar solution becomes less inevitable when the next country solves the same problem differently. That gives his later interest in comparison a human background: before he helped people inspect competing fares, he had lived with competing versions of ordinary life.

At Dartmouth, where he graduated in 1991, he studied economics and history. He took no computer science. Looking back, he valued learning to construct an argument, recognize capable people and communicate a possible solution. A technology company still needs someone to explain which problem deserves solving. His route into the business ran through that door.

An MBA at Northwestern's Kellogg School followed in 1997, along with consulting work in the years before his travel ventures. His subsequent career would depend heavily on technical colleagues, but his own account of preparation is strikingly verbal: define the problem, describe the answer, get people moving. The boarding pass came after the argument.

“If you’re not in a job that you’re passionate about, you’re not going to be thinking about it in the shower.”

Steve Hafner · 2016

A search box with fewer obligations

At Orbitz, Hafner was part of the early team behind the online travel agency launched in 2001. He helped build a business that grew from a handful of people to a substantial operation. By 2003, he was ready to leave. The experience supplied both knowledge of the industry and a complaint about the way it served customers.

Travelers wanted to compare what was available across the web. They might then prefer to purchase directly from the airline or another travel provider. Hafner saw an opening for a company organized around that search, rather than insisting on handling every booking itself. It was a modest alteration to the itinerary with large consequences for the business.

KAYAK, founded with English in 2004, made comparison its central job. Its usefulness came from bringing alternatives together so the customer could decide. For the traveler, the attraction was wonderfully unromantic: fewer separate searches before making an expensive choice. The holiday could supply the romance later.

The company went public in July 2012. Priceline announced an acquisition agreement that November, and the transaction closed on May 21, 2013. The buyer recorded a purchase value of $2.1 billion. Hafner stayed on as CEO. The sale changed the ownership of the business without ending his working relationship with the problem that had prompted it.

English left KAYAK's operating leadership in 2013. Their partnership had depended on mutual confidence across their different responsibilities: each could make decisions while the other was away. That is a practical kind of compatibility. A cofounder relationship earns its keep when the calendar, rather than a carefully arranged meeting, decides who must act.

The BlackBerry objection

Hafner's judgment did not always point toward the next interface. When English proposed building a BlackBerry app, Hafner doubted people would want to inspect airline tickets on a phone. Surely they would prefer a large screen. English pushed ahead. By 2017, mobile accounted for more than half of KAYAK's traffic.

It is an unusually useful founder anecdote because the wrong answer belongs to the chief executive. The product improved because his objection did not end the discussion. Having an opinion and giving it permanent authority are separate management decisions. A small screen was enough to expose the distance between them.

Speed became part of Hafner's stated approach to running the company. He wanted people to have room to decide and make mistakes, while avoiding long delays. His message to students at Kellogg in 2023 also stressed curiosity, approachability and the dangers of accumulated process. An organization can pack so much procedural luggage that departure becomes the day's main difficulty.

Those ideas help explain the appeal of starting with a small team again. The founder who once dismissed another startup was never especially attached to bureaucracy. Remaining at a company for years and wanting it to move quickly can coexist. The harder task is persuading a mature organization to preserve that room for movement.

An old name, a different reservation

Lola's name had been waiting longer than its current technology. Hafner and English had considered it before KAYAK launched, but the domain cost too much. English later used it for a separate travel business, whose technology and team went to Capital One in 2021. The name eventually found its way back to their shared work through Booking Holdings.

The new Lola is owned by Booking Holdings, with Hafner reporting to its CEO, Glenn Fogel. It gives the founders a fresh product within a company that already owns travel and dining businesses. At launch, the team numbered about twenty people in New York and Cambridge, Massachusetts. Small-company speed sits beside access to large-company relationships.

Hafner has been precise about the intended occasion. The focus is frequent experiences, including things people do in their home markets. In early October, he described hotels and restaurants as already available, with events and activities next and flights and car rentals on the roadmap. That sequence matters. A proposed capability deserves a different tense from a working one.

The underlying bet is that the useful assistant must get beyond an appealing suggestion. Live availability, current prices and a way to reserve the result make the conversation consequential. Direct connections to partners are central to that approach. An eloquent description of a fully booked restaurant still leaves a person looking for dinner.

For Hafner, the design also revisits an old question about how recommendations are ordered. In 2022, he questioned systems whose operators could not explain why they showed particular results. His concern was understandable to anyone who has stared at a list and wondered whose interests determined the first item. As search becomes conversational, that question follows it into the chat.

The bill arrives monthly

Lola's membership model makes that question commercial as well as technical. The service offers a free tier, with Plus at $5 a month and VIP at $25, both paid tiers carrying a twelve-month commitment. The company says it excludes advertising and paid placement from recommendations. Those are commitments the product will have to honor each time it suggests somewhere to go.

Hafner's stated intention is to build loyalty through personalization and pass benefits back to members. The proposition resembles a familiar exchange: pay for access, then use it often enough to value belonging. Whether someone needs that arrangement depends on the reservations, rates and assistance it delivers. A membership card, even a digital one, has limited entertainment value by itself.

There are people behind the assistant, too. Lola's partnership with Ten Group supplies concierge support to paid members. Hafner describes human expertise as useful when a request calls for an experienced person who knows whom to contact. For a founder working on AI, it is a practical acknowledgment that some reservations still depend on a conversation between humans.

Away from product strategy, his preferences are pleasingly specific. He has named Stockholm and its islands, St. Barts and the Greek islands among favored destinations. His chosen book is A Confederacy of Dunces, which he has said he tries to reread annually. A travel executive with a taste for that novel is at least equipped to recognize a plan going sideways.

Lola leaves Hafner with a familiar assignment in unfamiliar form. He is again working with English, again trying to remove an inconvenience people have learned to tolerate. The test now begins before the destination is chosen and continues until a reservation exists. After years spent helping people find a way out of town, he is attending to the evening they might have without leaving it.