Another batch of Silicon Valley startups takes on car retailing

Editor's note: The report has been corrected to accurately reflect flat fees charged by Roadster. Carvana's revenue has also been corrected.

SAN FRANCISCO -- With the rise of e-commerce, the Internet became a great place to shop for cars, but not a good place to buy them, despite the high hopes of a wave of startups that crashed on the rocks of the dot-com bust of late 2000.

Fast forward by 15 years, and Silicon Valley has come back to the dream of e-commerce for cars with a vengeance. According to CrunchBase, a website that tracks startup funding, four online marketplaces for used cars -- Beepi, Carvana, Shift and Vroom -- have raised $530 million since 2013, refining their technology to mount what could be the largest challenge to brick-and-mortar retailers.

New cars could be close behind.

San Francisco-based Roadster, formed in late 2013 by a group of e-commerce veterans who started the fashion shopping website ShopStyle, has an ambitious goal: to reinvent the new-car buying process. The point is not to get consumers into the dealership, but to avoid it, CEO Andy Moss said in an interview.

"We want to deliver a real e-commerce experience," Moss said, "not just a portal that brings you into the old business model."

Market research suggests many people, accustomed to online shopping, are ready to skip dealerships and buy cars online. According to a 2015 survey by the consultancy Accenture, 75 percent of consumers would consider doing an entire new-car deal online, from pricing to financing to paperwork, before taking delivery to their home.

Customers who visit Roadster's website see a bold slogan: "Skip the Dealership." Beepi, a Silicon Valley used-car startup known for putting bows on its cars before delivering them to buyers, uses a similar motto: "Find Deals. Not Dealers."

In the case of Roadster, there's still a brick-and-mortar showroom behind the scenes. In San Francisco, its pilot market, Roadster works with dealerships such as BMW of Santa Maria, which sees Roadster as a way to boost sales.

Steve Pangelina, the store's fleet sales manager, said he sells about 350 new BMWs annually across California to independent brokers and services such as Roadster. Every time, the store needs to load the car on a flatbed and ship it to the buyer.

"There's more work involved doing a sale like this, but the customer wants it," Pangelina said. "They don't want to go to a dealership, which is why they go through companies like Roadster in the first place."

For dealerships that make their money from financing, or by selling service, this approach may not be worth it. His store, nestled in an agricultural town 270 miles down the coast from San Francisco, doesn't have those options.

"But it has to be a good relationship," Pangelina said, adding: "Without the dealerships, Roadster's out of business."

A test drive at a dealership might seem essential, yet many customers now use the Internet to order eyeglasses from Warby Parker, mattresses from Casper and engagement rings from Blue Nile -- all without trying them first. The used-car startups have a seven- or 10-day return policy, but Roadster, which is ultimately facilitating deals with dealerships, does not.

"There are some directional moves in consumer behavior that make e-commerce better, but nothing magical has happened that has made e-commerce viable," says Ernie Garcia, CEO of Phoenix-based Carvana, which has 2,500 used cars in its inventory and and claims to have brought in $140 million in revenue in 2015. "It's that the platforms are finally good enough to make it worthwhile to buy a car online."

That was the problem in the 1990s. In the early days of the Internet, dozens of startups tried to sell new cars online, only to find that the marketplace wasn't mature enough for the task and that state franchise laws restricted their options.

A few of these startups survived, including Autobytel.com and CarsDirect.com, but by selling leads to dealerships instead of selling cars to customers.

Even the most successful car-shopping startup of the 2000s, TrueCar, has struggled to regain its footing since 2012, when disgruntled dealers ditched the service in droves, claiming TrueCar eroded transaction prices and margins. The company lost $65 million on $260 million in revenue in 2015 as founder Scott Painter stepped aside as CEO and AutoNation Inc., investing heavily in its own digital storefronts, stopped buying TrueCar's leads.

Roadster's strategy is different: to be a 21st-century broker.

Since the early days of the automobile, when entrepreneurs such as Henry Ford wooed wealthy families with the promise of a vehicle better than a horse-drawn carriage, there have been brokers like Christopher Gross.

Gross, a corporate finance veteran with an MBA from Harvard Business School, ran a company called BuySide Auto in San Francisco's posh Pacific Heights neighborhood, home to much of the city's political and business elite. Nancy Pelosi, the longtime Democratic leader of the U.S. House of Representatives, has a red brick mansion there. Larry Ellison, co-founder of Oracle Corp. and the third-richest man in America with a net worth of $47.5 billion, lives nearby on a block called Billionaires' Row.

His customers, mostly well-off professionals with more money than time, would pick out a car, and Gross would do the work. He would scan dealer inventories, run the numbers on a loan or a lease, and use his dealer contacts to get a good deal. Customers would pick up the car in Pacific Heights, or pay extra for delivery.

It was e-commerce without a website: You ordered a car from your house, and it showed up there, as if by magic. Gross was selling 25 cars a month and charging up to $795 for each car he sold -- a "great little one-man business," he said.

Yet the success of the business relied on Gross' finance skills, and his relationships with dealerships. It relied on his knack for reassuring his customers, many of them friends and friends of friends, who were worried about getting the best deal.

"It doesn't scale well," Gross says. "It's a very high-touch business."

At the end of 2014, Gross tried something new. He sold his business to Roadster, which set off on a mission: to turn Christopher Gross into a product.

Since then, Roadster has aggregated dealer inventories across California, refined its pricing methods by collecting data from third-party sites such as Kelley Blue Book and developed a way for customers to find a car with an exact set of features.

Roadster charges a $295 flat fee for a service called Concierge offered only in the Bay Area. It's a digital broker business -- behind the Web page, there are people like Gross making calls and making the deal happen.

"That is not the business we want to be in," Gross says. "That's the business I wanted to be in [with BuySide Auto]. But that's not the business where you're going to grow and do transactions across the country at a rate of thousands a week."

Laws for brokers vary from state to state, which could slow or impede the spread of Roadster's business model. California, for instance, requires brokers to be licensed, to have a physical place of business and to make delivery through a licensed dealership, says Brian Maas, president of the California New Car Dealers Association.

The association has sued TrueCar, claiming that it requires a broker license. The group hasn't challenged any of the new e-commerce startups, Maas says.

"There are lots of brokers operating throughout the state, and as long as they're following the law, we're OK with it," he says. "If they're brokering, they're brokering -- it doesn't really matter if they have a website with a dot-com behind their name."

Roadster, which handles about 100 transactions per month, advertises on local AM/FM radio, which is enough to get the word out, but not so much that its staff gets deluged.

This month, Roadster took the next step by launching Express, which takes the human broker out of the equation. For a $295 flat fee paid by a customer, Roadster will help customers buy one of the 100 most popular vehicles and take delivery at home. Limiting the service to the most popular vehicles makes it easier to set prices, because prices for less-common vehicles are less predictable. (Roadster would probably charge less for Express than for Concierge service once the services mature a bit more.)

This differs from TrueCar. Instead of giving customers a price guarantee that dealers are supposed to accept if they want TrueCar's leads, Roadster makes offers to dealerships, asking if they want a given deal. Instead of charging a flat fee, as TrueCar does, Roadster takes a cut of the transaction worth up to 1.25 percent from the dealer.

Roadster COO Rudi Thun says the company is trying to build an e-commerce platform that works for the next 15 or 20 years. It could also be licensed to automakers as an online selling channel that helps customers pick out their cars and conveniently fill out paperwork before they step foot in the local dealership.

As it grows, Roadster will need to go city by city and state by state, dodging state laws against brokers that would curtail its business. Thun sees that as an obstacle, but not a deal breaker, for the company's strategy. "We can work within the system," Thun said. "And if it changes, we'll be fine."

Like any San Francisco startup worth its salt, Roadster already has a crosstown rival. Based a few miles away, Drive Motors has raised $1.5 million in capital from blue-chip venture capital firms such as Khosla Ventures and has landed a spot at the prestigious Mountain View, Calif., startup incubator Y Combinator.

Drive Motors, which plans to set itself apart with dealerships by charging them a flat monthly fee instead of taking a cut of each sale, has signed up 100 dealerships in California, Aaron Krane, founder and CEO of Drive Motors, said in an interview.

Krane said he came up with the idea for Drive Motors late at night while sitting at home watching a basketball game. Flush with cash from having sold his last startup, Hitpost, to Yahoo, Krane decided he wanted a new Mercedes-Benz CLA45 AMG.

He discovered he couldn't order the car online. He didn't want to visit a dealership. So he waited. By the time he bought a car, he had changed his mind.

"Sucks for Mercedes," Krane said. "Now I'm a BMW owner."

Automakers are determined not to lose young customers such as Krane because they don't offer a quick-and-easy digital buying experience.

For that reason, General Motors has launched Shop-Click-Drive and the Factory Pre-Owned Collection, which allow customers to start buying a new or used vehicle online, though they finish the process at a dealership. Toyota, which built a similar shopping platform for Scion called Pure Process Plus, is now working to bring it to Toyota.

The trend toward e-commerce has also caught the attention of retailers such as AutoNation and CarMax, which could use their national reach and healthy brick-and-mortar retail businesses to build e-commerce brands of their own.

At the end of 2014, AutoNation committed to invest $100 million in what it calls digital storefronts.

Customers choose a car online and get a certificate, then place a deposit on PayPal using a credit card. With a car reserved, the customer then visits the dealership to take a test drive and finish remaining paperwork.

AutoNation's goal is to get customers out of the store in 30 minutes or less.

"We see a frustrated consumer that currently has an online experience and an in-store experience that are very much disconnected," AutoNation CEO Mike Jackson told Automotive News when the project began. "We see that as an opportunity."

Still, a $100 million investment is a small drop in the bucket compared with what AutoNation, which earned $3.1 billion in gross profit in 2015, spends on brick-and-mortar retailing. Heavily invested in selling cars through showrooms, the company is more interested in displacing third parties such as TrueCar than in upending its own business model.

Yet if startups win a significant share of car sales, retail giants such as AutoNation will be compelled to snap into action, either acquiring or imitating them.

The history of e-commerce shows that it's important to be prepared well before that day arrives. Wal-Mart, slow to heed the ascendancy of Amazon in the 1990s and 2000s, has since spent billions of dollars to become a legitimate rival in online shopping.

Mark Rikess, president of Los Angeles retail consultancy Rikess Group, says smaller retailers may struggle with e-commerce, but the larger ones seem to be open-minded enough to avoid playing the role of Wal-Mart to the newcomers' Amazon.

"There is certainly room for additional players," Rikess says, "but if you show CarMax a way to do things better than they do right now, they're going to do it."