Langley Steinert and the art of being wrong

Langley Steinert had an idea for your holiday. Answer a series of questions, let the software consider the possibilities, and receive a recommendation for where to go. VacationAdvisor sounded useful to him. Six months of development later, the customers were considerably less excited.

He wanted to keep improving it. Then he and his co-founder looked at the usage and stopped. The feature had acquired something more valuable than polish: an answer. People scarcely used it. Continuing would make the mistake more expensive.

Steinert would go on to explain this failure in his own writing. There is something disarming about a founder who supplies the cautionary tale himself. The embarrassing product, the wasted months, the instinct to double down: he put all of it on the record. His advice had a receipt attached.

The holiday nobody asked for

By then, he had already learned that the name on a company’s door need not describe the business that eventually pays the rent. TripAdvisor started in 2000 with a search technology and a plan to license it to other publishers. Steinert was chairman and co-founder; Steve Kaufer was CEO. The technology could locate travel information. Finding somebody willing to pay for it proved harder.

With cash running short, the founders reconsidered the audience gathering at their own site. They could build a consumer business and earn money by referring people to booking services. A test with Expedia helped establish a workable revenue model. The supposed demonstration site had become the business.

The irony is appealing. A company built to help people find their way had needed directions itself. Its founders had to distinguish between interest in the product and demand for the arrangement they wanted to sell.

“There can be no sacred cows in the world of start-ups.”

Langley Steinert, 2012

Before the reviews, the apprenticeship

He had not arrived in software by the direct route. After Georgetown, Steinert entered investment banking in 1985. The long hours helped persuade him to find another line of work. An MBA at Dartmouth’s Tuck School of Business, completed in 1991, became a bridge into technology.

His early software career included Lotus, Papyrus, and Viaweb. At Viaweb he worked in marketing and business development before Yahoo acquired the company in 1998. These were jobs inside companies making and selling products, years before his own name became associated with consumer marketplaces.

The apprenticeship left him with a particular impatience. He remembered enterprise software development cycles lasting 18 months. In a web business, he argued, waiting that long postponed the most useful conversation: the one with the people actually using the product.

In 2012 he described CarGurus releasing features nightly. A simple email service for relevant car deals went out with limited functionality; customer response justified further work. That was the inexpensive counterpart to VacationAdvisor. Give an idea enough substance to be tested, then let its audience decide whether it deserves another month of engineering.

The car needed a different question

Steinert founded CarGurus in 2006. The first proposition borrowed the familiar travel formula: people would review cars, exchange information, and help one another. There were discussion boards and a wiki-style approach to automotive articles. It attracted some users, without producing the business he wanted.

This time, previous success could be particularly misleading. A review site had worked in travel. Why should a review site for cars be such a stretch? The analogy was tidy. The results were untidy enough to require a change.

Steinert later acknowledged another handicap. At TripAdvisor, his business experience had complemented Kaufer’s technical and product interests. He started CarGurus alone and felt the early service lacked a clear enough product vision. Bringing in a strong engineering leader helped. He had learned something about the limits of doing the founding himself.

For its first three years, CarGurus was Steinert and six developers in a Harvard Square townhouse. A small team meant fewer salaries to cover while the product found its purpose. He treated a low spending rate as permission to make mistakes and survive them. The office arrangement was modest; the room it gave the company to reconsider its direction mattered more than the square footage.

CarGurus moved toward a shopping question with a more immediate consequence: which available car represented a good deal? A prospective owner might appreciate an article about a model, but still need help judging the vehicle for sale down the road.

A price tag needs company

A seller can give you a price without giving you much context. CarGurus set out to supply the missing comparison. Its Instant Market Value product, launched in 2010, helped turn vehicle data into an assessment a shopper could understand.

Steinert’s account of the search approach included price relative to the market and the dealer’s reputation. A cheap car and a dependable seller are separate considerations. The product tried to put both into the decision, alongside information such as price changes and time advertised.

For dealers, that scrutiny could be uncomfortable. In a 2016 video interview, Steinert faced questions about whether the ratings could mislead shoppers. His argument was commercial: buyers arriving with an outside assessment of value could feel more comfortable proceeding. He presented transparency as something that could help sell a car.

There was a negotiation here, rather than a fairy tale about everybody wanting precisely the same thing. The shopper wanted a sound purchase. The dealer wanted a worthwhile customer. Steinert needed enough of each to keep the marketplace useful.

When CarGurus expanded to Britain in 2015, he emphasized building its audience before pushing revenue. The American operation could support the investment. He also made bold predictions about catching the incumbent. By the following June, he conceded that one inventory target might have been aggressive. Even the forecasts were subject to revision.

The chief executive takes the call

The public milestones can obscure the ordinary jobs Steinert liked to discuss. Once, when somebody trying to schedule a meeting asked for his executive assistant’s contact details, he answered: “You’re talking to him…” There was no assistant to transfer the call to.

At CarGurus, he took on prospective sales accounts himself during a new initiative. He did not close every lead. He did close one deal, and the experience gave him a closer understanding of what his salespeople were facing. A rejected sales pitch contains information a management presentation may politely omit.

He had seen a version of this at TripAdvisor, where Kaufer checked the product and sent developers detailed messages about problems he found. Seniority could coexist with the unglamorous business of noticing something broken.

Steinert’s account of office life had little affection for elaborate meetings. Developers had considerable freedom over functionality and interface decisions. In 2014, he described taking five randomly chosen employees to lunch twice a month and asking what should improve, from meals to product direction. He promised to listen, without promising to implement every suggestion.

Lunch offered a practical test of accessibility. Anybody can declare an open door. Sitting down with people selected without regard to their rank creates a different chance of hearing something inconvenient.

“Leadership means setting an example.”

Langley Steinert, 2014

The people who owned a piece

CarGurus went public in October 2017. Steinert’s explanation of that decision began with a private-company plan that had become insufficient. He had imagined a profitable business distributing money to shareholders. Recruiting employees with equity brought another obligation: a route for those employees to realize its value.

There was a governance condition attached. CarGurus used a dual class voting structure that preserved his control. The episode combines employee liquidity with a founder’s desire to keep steering the company. Both belong in the account.

His preference for lower early spending also had consequences beyond frugality. Founder Collective described an early CarGurus customer accounting for a large share of revenue, then ending the relationship as the company tried to diversify. The sales team rebuilt the business beyond the previous revenue level within a year. Depending on fewer promises from a single customer made the company less exposed to that customer’s next decision.

In January 2021, Jason Trevisan succeeded Steinert as CEO. Steinert became Executive Chairman. The 2026 company filing still identifies him in that role, alongside his chairmanship of the board. A change of job had left him connected to the business he founded.

Another address, familiar colleagues

ApartmentAdvisor extended the sequence from trips and cars to places to live. Steinert became its co-founder and board chairman in October 2020. Its early team included people with experience at the companies he already knew.

The venture now reaches beyond apartment listings. Its ASSIST software addresses property management work including leasing, maintenance, and payments. In October 2025, ApartmentAdvisor announced an integration with Rent Manager. The work has moved from helping somebody find a home into parts of the operation that follow.

Across these businesses, the recurring interest is legible: people trying to make consequential choices, and software that can provide context or remove a troublesome step. The industries differ. So do the products that eventually earn their place.

He has also offered a less technical explanation for continuing to build. Speaking to his children’s school about TripAdvisor, he was asked by a seventh grader whether he had made much money. Steinert laughed. He described enjoying the work itself: the questions, the developers, the chance to make a product. An exit could settle the financial question without settling what he wanted to spend his time doing.

VacationAdvisor remains a useful footnote to that career. It was an idea its founder liked, supported by months of work, and rejected by the people meant to use it. Steinert could tell the story because he had stopped building. Sometimes a company’s next useful product begins with a founder finally agreeing to leave the previous one alone.