How Traditional Automakers Can Still Win Against Google's Driverless Car: Part 5 (of 7)

11 years, 6 months ago - 18 April 2013, Chunka Mui, Forbes

Cars are such a mature technology that we’ll soon reach the 250th anniversary of the first automotive accident—a Frenchman who rigged up an early steam engine to pull artillery, lost control and ran into a wall in 1771.

As a result, automakers, even more than most companies, are sure they know the trajectory of their industry and who their competitors are. After all, for almost a century, pretty much the same set of companies has been going at each other, hammer and tongs—plus, nowadays, power socket wrenches and industrial robots.

The temptation will be to dismiss the Google driverless car as pie in the sky. Google may have plenty of intellectual horsepower, but there’s no vroom, vroom to what Google does. Google doesn’t know anything about how to build cars, what car buyers want, etc., or so the rationalizations will go. Far from being car guys, Google distributes old-style bicycles around its campus in Mountain View, CA, for employees to grab at will and ride to other buildings.

If the car companies dismiss Google’s efforts, they risk missing the biggest transition their industry has seen since Henry Ford’s automated assembly line.

If, however, the car companies approach the transition correctly, weighing all the threats as well as all the opportunities, the companies should win. They have access to essentially the same driverless technology as Google, while also possessing hard-earned expertise on stamping metal, building engines and managing complex supply chains. The car companies have trusted brands, loyal customers and years of data on the cars themselves and on how customers buy and use cars.

The Automakers’ Dilemma

But, just because the car companies should win, doesn’t mean they will. Plenty of market leaders have missed cusps and fallen away. Think Kodak, Borders, Blockbuster, music companies, newspapers… I, personally, sat in the office of the CEO of one of largest newspaper conglomerates in 1997 and argued that the Internet could kill his business, but he argued that the company would be fine. By the mid-2000s, investors had forced the sale and dismantling of his enterprise.

For the car companies to win, they need to at least contemplate major disruptions to their mindsets, organizational structures and business models, but this will be a challenge. Many have just gone through near-death experiences and surely don’t want to contemplate more. The companies feel like they have solved their biggest problems, related to union contracts and debt.

Besides, automakers are filled with “car guys.” Even the women are car guys. Car guys love cars and the thrill of driving. Some still have a difficult time accepting automatic transmissions, anti-lock brakes and cruise control. Imagine their emotional reaction to not driving. The car-guy mindset is so strong that it could limit the possibilities for innovation that the incumbents consider.

Organizational structure may be an issue because companies that, like the automakers, have a strong central organization or that are organized by function tend to have rigid processes and emphasize efficient operations rather than innovation. Such companies tend to keep doing what they’ve always been doing, just a little better each quarter, rather than look for breakthroughs.

More than any other factor, business model considerations will limit automakers’ ability to pursue the potentially disruptive innovation consequences of the driverless car. What car company would try to invent a future that included shared usage of significantly fewer cars? That future would fly in the face of business models that depend on personal ownership of cars sold through a powerful network of car dealers.

The most likely course will be for automakers to continue on a very incremental approach. They will develop impressive technology but will work towards adding features for intelligent driver assistance and resist jumping to driverless operation of cars. This approach is completely rational in the short term, and will not disrupt mindsets, organizational structures or business models.

That means automakers would be slow in dealing with all the disruptive implications that I’ve laid out in earlier pieces in this series—perhaps 90 percent fewer cars, many of which will be shared rather than individually owned; automated operation, with the human mostly out of the driving loop; shorter, hits-based product cycles; almost no accidents, meaning far less need for heavy structural elements and for safety features; and so on.

Fully automated, driverless technology may, in fact, ultimately be adopted incrementally, but an assumption of incrementalism will be very limiting and open the door for upstarts willing to contemplate radical innovation.

Even if there is only a 10 percent chance that the future will be as disruptive as I have portrayed (I believe that odds are much higher), automakers need to make sure that they are the ones that invent it, as opposed to assuming (or hoping) that others will not.

The car makers need to Think Big, Start Small and Learn Fast, as Google is doing. Here’s how they should proceed:

Think Bigger

Strong cultures tend to put a pillow over disruptive ideas and smother them before they get to be big enough to defend themselves.

For instance, the U.S. Postal Service did a study some years back on whether digitalization was a threat and mostly dismissed it. Forecasts were so over-the-top optimistic that the worst-case scenario for first-class mail four years from now was 10 percent higher than volumes are now.

And don’t just say the USPS is full of fools. The mistake they made is quite common among incumbents during technology-driven upheavals.

The first step to thinking bigger is to set aside what one expects or hopes to happen and go looking for doomsday scenarios.

The fundamental challenge poised by driverless technology to automakers is that it enables a shift to a shared usage model for cars while dramatically improving cost and the customer experience. Automakers need to understand all the reasons why the new model could be a deadly scenario for them.

First, according to Larry Burns, former head of research, development and strategic planning at General Motors, cars in a shared-usage model will evolve into dramatically new forms. For example, fleet-level car sharing allows for matching the vehicle to the particular trip. Because most car trips are short ones involving one or two people, most cars in a shared fleet could be smaller, simpler and cheaper electric vehicles—much like the ENV concept car that GM unveiled a few years ago:

Customers will no longer need to buy based on their worst-case scenario: the longest trip they might take, with the largest number of kids in the soccer car pool.

Burns, who now holds teaching appointments at the University of Michigan’s College of Engineering and Columbia University’s Earth Institute, says cars like the cheaper and easier-to-make ENV-like cars will shift customer demand toward commodity products.

Such commodity cars would erode already thin profit margins and undercut business models that depend on the allure of premium autos and on options that carry large profit margins. Simplicity in cars would also reduce the traditional competitive advantages that stem from automakers’ ability to manage complex supply chains and integrate thousands of parts. Simpler cars would also reduce the necessary engineering expertise and capital costs, two barriers to entry that have protected automakers from new entrants.

The shared-usage scenario also threatens to turn other traditional automaker assets into liabilities. For instance, dealer-based distribution systems have little value if individual ownership fades. Automakers need to ask themselves how dealers might try to hold them back.

Given the historically contentious relationships with unions, automakers also need to ask how labor issues might hinder their ability to adopt simpler car designs and radically different manufacturing processes.

Another difficult aspect of shared usage is that it upends automakers’ traditional planning and business cycles. If all our driving is crammed into 10, 20 or even 50 percent as many cars, those cars will be used up and replaced much faster. Rather than having the population of cars turn over about every 15 years, it might turn over every couple of years. Cars would become much more of a hits-based business, and it would be crucial not to miss a product cycle—look at what happened to Nokia, Palm and RIM when they missed a cycle in cellphones.

Even worse, what if Google extends the cellphone analogy and offers its driverless OS to manufacturers for free, to have people use its search capabilities in cars, its maps, its business software and so on? What if Google or some other company borrows from the cellphone industry and greatly subsidizes the purchase price of a car, in return for a two-year contract to use certain services? How could a car company compete with free or even almost-free?

There are plenty of other possible doomsday scenarios, too, based on developments with electric cars, with regulation, with competitors. What happens, for instance, if the emerging middle class in China somehow generates the modern-day equivalent of the Volkswagen Beetle and sweeps through the US market?

Once the automakers truly accept the possibility of their doom, they can move on to the second part of thinking bigger. They can go on offense.

The companies need to get out a clean sheet of paper and reimagine what their business could become in the context of driverless cars. It’s important not to think about current constraints, such as liability issues, the talent currently within the company, the existing organizational structure or the way that current car designs could incrementally change. The how-to-get-there-from-here stuff will have to be dealt with, in time, but focusing on it too early makes it too hard to truly Think Big.

Automakers, whose traditional business model is built around supplying cars to car dealers, might use a clean sheet of paper to lay out business models where they engage more directly with consumers and share in the multitrillion-dollar revenue stream associated with cars.

As I described in Part Two of this series, driverless cars have the potential of dramatically reducing the transaction costs associated with transportation, including the capital cost, depreciation, fueling, maintenance and insurance.

Driverless cars might also spark enormous new revenue opportunities to serve connected vehicles and their captive occupants, now freed from the task of driving. So, automakers should imagine how to put themselves smack in the middle of all the value in thecar market, rather than building cars on razor-thin margins while letting others reap most of the profits.

Assuming that personal ownership will persist, in some part, for a while, the companies might conceive of themselves as a concierge service around the maintenance of the car and other kinds of activities customers associate with driving. They might help coordinate any service or repairs. Taking a lead from the utility market, where OPower uses data from utilities to let consumers know how their energy use stacks up against comparable households, the car companies could let customers know how their fuel efficiency compares with other owners of the same type of vehicle, then suggest ways to improve.

Or, automakers might help car owners earn extra money by renting their unused cars out to strangers—after all, cars are only in use about 5 percent of the time; why have such an expensive asset just sitting around? General Motors has already set up a partnership with RelayRides, a company that facilitates “peer-to-peer carsharing,” where car owners rent out their cars by the hour. Using OnStar, GM car owners who sign up with RelayRides can leave the keys in the car and have renters unlock it by using a code on a smartphone, thereby saving the hassle of exchanging keys in person. For RelayRides, the deal makes it immediately easier for more than six million OnStar subscribers to participate in its program. For GM, the deal attracts young and urban users who might not otherwise consider GM.

Maybe companies would reconceive cars as a platform like iPhones and let third parties create apps that could be sold through something akin to the Apple App Store.

A friend who works at a big car company and tinkers offered some possibilities based on bits of software he’s written using his access to the company’s application programming interfaces. He has his car text his wife when he’s 10 minutes from home. (He did once cause some confusion when he went out of town and lent his car to a friend who lived nearby. As the man returned home from work, to a house a few streets over from my friend’s, the car dutifully texted my friend’s wife. Confused, she called my friend and said, “What’s going on? I thought you were out of town.”) When his teen-aged daughter takes his car out at night, he has it give her a 10-minute warning, based on her location, about when she’ll have to leave to make it home before her curfew; he says the notification makes her feel more secure. He says an electric car could be programmed so that it would only draw power when rates fell below a certain level. He’s thinking of having his wife’s car text him when it’s down to a quarter-tank, so he can fill it up.

Automakers could imagine a world with far fewer dealers, where people can go, see and test-drive cars, then order them directly from the manufacturer. Perhaps the manufacturers would conceive of dealerships as like Apple or Zara stores: hip places where people would visit frequently to buy the latest cool apps for their cars.

Moving into the driverless future, automakers could reimagine one of their scariest concerns about driverless cars—liability. Automaker fear that taking the driver out of the loop will shift liability for accidents to them. But in a world of virtually no accidents, where cars can’t be stolen, where fault is easily determined via onboard black boxes, and where the risk profile of trips can be managed through knowledge of speed, routes, time-of-day, etc., might it make sense to embrace such liability? Hundreds of billions of dollars are spent each year on personal and business auto insurance. Perhaps automakers could aggressively reconceptualize that business themselves.

Car companies might also could imagine an opportunity to be an intermediary managing car-sharing. People would want assurances that shared cars are well-treated and maintained—including having the latest driving software—and automakers bring many assets to that role.

Thinking big means that automakers must contemplate plausible futures where their worst fears about driverless cars come true, then leverage their creativity and assets to conceptualize business models that turn doom into boom.

Start Small

Having thought big, the companies must resist the tendency to immediately bet big on their ideas. Instead, companies need to move into a testing phase where they Start Small and Learn Fast. The tendency, especially in huge companies with huge planning departments, is to try to build long-term plans based on premature analysis. The result is grand strategies based on guesses masquerading as fact.

The car companies, instead, need to focus on gathering real data through a series of small tests for the scenarios that seem to hold promise. As I’ve said, the Google car was sparked by the early work of just a dozen engineers, so even tests of radical ideas can be relatively small and inexpensive.

Before laying out those tests, automakers must ask themselves hard questions about the breadth of the technology capabilities they are currently developing for a driverless future.

Some industry experts have told me that some automakers, especially the luxury manufacturers, have capabilities like Google’s under wraps in their laboratories.

Other industry insiders are skeptical. They say that while Google uses readily available components, its driverless software is unparalleled. They say most manufacturers and Tier 1 auto suppliers made two critical assumptions that limited their technology work. They assumed that an incremental, driver-assistance approach was the way to go. They also assumed that vehicles would be linked to each other and a supporting infrastructure, as opposed to Google’s model of relatively stand-alone vehicles (communicating, not surprisingly, mostly with other Google resources).

Whether the skeptics or optimists are right—and the assessment will vary by manufacturer—every organization must have a clear-eyed view of where it stands. The possibilities for fully driverless, Google-like cars are too important not to have a strong IP portfolio—or a vigorous strategy to build one.

After delving into the technology issues, companies that are ready to start small could begin with tests similar to what I proposed that Google should do to accelerate its own driverless program.

Google has adroitly developed its technology, burnished its brand and captured public imagination with extensive road testing of its driverless fleet. With recent legislative changes that have allowed the car on the roads in California, Nevada and Florida, Google is likely to expand those efforts. In the process, it will log millions of road miles, generate mountains of data on the safety and benefits of the car and crack lingering problems. For those automakers that are able, there is no reason to cede this high-profile public relations and learning window to Google. Car makers should adopt similar programs.

Companies should also consider forming a partnership with Google. It doesn’t want to make cars, so there could be a way to divvy up responsibilities and opportunities in ways that would work for both sides. If you look at how much AT&T benefited from getting early, exclusive rights to the iPhone, you can imagine the benefits that a partnership with Google might provide a car company.

Automakers could establish pilots that illuminate and test aspects of disruptive business models coming out of their Think Big efforts. An automaker could, for example, take the lead in the Big Venture Play scenario that I described in Part 3 of this series and launch a shared, driverless transportation system as a test. A detailed analysis of Ann Arbor, MI, concluded that a shared-driverless system could be fielded that offered customers about 90 percent savings compared with the cost of personal car ownership, while delivering better user experiences. It would be great to have real data.

Car makers are already trying some novel things. For instance, GM recently formed a venture with AT&T that could turn cars into WiFi hot spots. But car makers should experiment more broadly, perhaps using Nike as an example.

Nike sees how sensors and mobile devices could transform healthcare by letting people stream information on their bodies to a doctor or other analyst, including a computer. Nike also realizes that people who buy its sports equipment and clothing could be the early adopters of health-related technologies. So it is experimenting with a broad array of possibilities for embedding health-related sensors into its products. Automakers need to be sure their tests run the gamut and are as forward-thinking as Nike’s.

To reiterate: The objective is to take a broad, inexpensive, experimentally driven approach to eliminate uncertainty and start understanding how the future really could look.

Learn Faster

When I asked Larry Burns what advice he would give to his former colleagues in the automotive industry about driverless cars, he replied, “It’s all about the learning cycles.”

He cautioned that industries faced with the potential disruptions of driverless technology should not move too quickly to a fixed conception, because the future is still very uncertain. Instead, companies need to establish footholds that enable them to learn and learn and learn—then learn some more.

While he, like many others I talked to, believed that the technology was inevitable and could become viable sooner than most think, he felt that the window of opportunity was still open for automakers to leverage their assets and assert themselves in the coming transformation.

An invaluable benefit of undertaking something like the Big Venture Play is that it would drive tremendous learning around a whole range of issues, including technology, regulatory issues, customer behavior and preference, business models and venture partnerships. Because there is so much complexity to driverless technology and business models, companies need to be aggressive to make sure they don’t fall behind on the learning cycles. It will be very hard to catch up.

Burns drew a parallel between driverless cars and the level of disruption that digital photography brought on Kodak (as I have done, separately). He hoped that automakers’ management could resolve the mindset, organizational-structure and business-model challenges that Kodak could not, and in the process not only survive but thrive in the transformed markets.

* * *

My friend and colleague Alan Kay has famously said that “the best way to predict the future is to invent it.” More recently, he said that many incumbent companies, like some in the car business, seem to think that “the best way to predict the future is to prevent it.”

If the car companies, however unintentionally, take an approach that would prevent the driverless car, or other fundamental advances, they will lose. If, however, they can Think Big, Start Small and Learn Fast, they can invent and own an extraordinary future in which culture, commerce and society reshape around driverless cars, much as those aspects of our current lives were shaped around human-driven cars.