Google's Trillion-Dollar Driverless Car -- Part 4 (of 7): How Google Wins

11 years, 8 months ago - 19 February 2013, Chunka Mui, Forbes

Driverless cars have the potential to save millions of lives and throw trillions of dollars in existing revenue up for grabs while sending a tsunami of business disruption across multiple industries.

But, what advantage does Google have over traditional car makers, some of whom have been exploring driverless technology longer than Google has existed? Even if Google succeeds from a research standpoint, how can it profit from efforts that are so far from its core business of search-driven advertising? These are the key questions I will explore in this article.

The question of Google’s strategic advantage in driverless cars has stirred up a hot debate in the virtual corridors here at Forbes. As Joann Muller and Haydn Shaughnessy have pointed out in their responses to this series, car makers have their own driverless car programs and clear plans for deployment. Joann and Haydn’s view of the prospects of Google’s winning over car makers are “baloney” and “nonsense,” respectively.

I confess that I am not as confident about a Google victory as Joann and Haydn are sure of its defeat. Google’s driverless cars might still hit a lot of speed bumps, and they might even crash into a wall. But, guided by research that Paul Carroll and I have done on thousands of corporate disruptive innovation efforts—including both successes and failures—I believe that Google is doing a lot of things right with its driverless car.

Google’s approach has already vaulted it into contention in a race that no one expected it to even be in. It also positions Google to profit handsomely. Here’s why:

1. Google is thinking bigger.

Alan Kay observed that “Point of view is worth 80 IQ points.” His point is that being smart doesn’t depend on just mastering knowledge or tools. Instead, the level of insight about how those tools fit into the world can make a big difference.

Our research shows successful innovators dare to Think Big. They focus on the killer apps that can rewrite the rules of a company or industry, rather than just looking for incrementally faster, better, and cheaper change. Thinking big allows innovators to start with a clean sheet of paper and consider a full range of design approaches and possible futures. They consider not only building on current capabilities and business models, but also moving in brand new directions.

Google thought big when it aimed to reduce accidents, the time and energy wasted in traffic, and the number of cars on the road by 90 percent. To attain such high aspirations, Google set an aggressive research and development agenda that employed more expensive technology and aimed for full automation, rather than using off-the-shelf technology that enable narrow applications.

Human error accounts for between 77 and 90 percent of all road accidents. Google’s long-term approach is to get drivers out of the loop, rather than make them better drivers. Also, car sharing is much more limited without full automation, meaning that a 10x savings in transportation costs and massive reductions in the number of cars would not be possible.

Google also thought big by recruiting world-class scientists and engineers, including Sebastian Thrun, Chris Urmson and Astro Teller.

Contrast Google’s approach with that of car makers, which are focusing on decidedly incremental approaches. Major car makers view driverless technologies as enablers to enhance the current driving paradigm. They plan to gradually introduce application-specific functions like Mercedes’ adaptive cruise control with steering assistance, Volvo’s Traffic Jam Assistance, GM’s “Super Cruise” system, and Audi’s piloted parking, where the car takes over only under well-defined circumstances.

From the car makers’ perspective, an incremental approach makes all the sense in the world. Introducing driverless technology bit by bit gives companies a long stream of premium-priced safety features that are completely consistent with current designs. A piecemeal approach also eases the specter of manufacturer liability issues that would arise if human drivers were not in the loop in terms of controlling the car but were nominally in control. Most importantly, the approach does not challenge existing business models.

Our research warns, though, about the dangers of thinking small in the face of disruptive technology.

Kodak, for example, failed because it allowed core patents in digital photography to languish in its vaults for decades. When Kodak roused itself to leverage its digital assets, it wasted years and hundreds of millions of dollars deploying them in incremental fashion—such as with its Advantix system, which used digital as a preview mechanism in a film camera.

It is human nature to see change as incremental and to think that customers want that, too. But, as Kodak, Blockbuster, Borders and scores of other former market leaders have learned, incremental thinking can leave huge openings for bolder companies willing to pursue new killer apps—like Google.

2. Google is starting small

Successful companies Start Small after Thinking Big. Rather than jumping on the bandwagon for one particular approach, they break the idea down into smaller pieces for testing. They defer important decisions until they have real data. They hedge their bets by making small investments in a variety of approaches, to see which work and which do not. In addition, successful innovators take the time to make sure that everyone—the executive team, employees, partners, regulators and maybe even customers—are working in unison, rather than having people pay lip service to a vision while actually working at cross purposes.

The Google car is the work of a mere 12 engineers, and the company has spent perhaps $50 million on the project. To put this amount into context, it is less than .0003 percent of Google’s revenue over the course of the program. It is also less than a third of what car makers have spent on Super Bowl ads over the same period.

Beyond limiting spending on technology issues, Google started small in its financial analysis. Early analysis is often premature—just guesswork masquerading as numbers—and there’s no way a typical business innovation process would have supported Google’s entry into the driverless car race. Google management, however, made the simple decision that it was playing in the right sandbox with the right people, and that was enough.

Google has also done a masterful job (thus far) of working through regulatory issues and wooing consumers and policy makers—by most measures, regulatory and consumer acceptance will be the real limiters to adoption. Google has been smart to work on these issues simultaneously with the technology development, rather than waiting until the technology was done.

Car makers are, of course, starting small as well. But that’s the best you can do when you are thinking small.

The danger of thinking small is that, oftentimes, companies that should be innovating in the face of a disruptive technology do not start at all. Our research found that companies that thought small tended to swing from complacency to panic. They think incrementally for too long and, suddenly finding themselves late to the game, make big bets on a single idea, only to have it not pan out. This is what killed Blockbuster, which ignored Netflix’s DVDs-by-mail model for years, then bet big on its own version before fully working out the economic and operational implications—Blockbuster’s business model didn’t work without the hundreds of millions of dollars of late fees each year, but management didn’t realize that until after it promised to halt the hated charges.

3. Google is learning faster.

Companies that Learn Fast take a scientific approach to innovation. They conduct extensive, inexpensive prototyping before they even get to the pilot phase—let alone the big rollout—so they can gather comprehensive information about their attempts at innovation and quickly analyze both what’s working and what isn’t. The successes also develop the institutional discipline to set aside or alter projects as soon as it’s clear that they’re not working.

Google is following Gordon Bell’s admonition that “a demo is worth a thousand pages of a business plan.” Rather than just researching driverless technology components or prematurely analyzing the potential return, Google is prototyping. By assembling a fleet of working cars and logging hundreds of thousands of miles in real road traffic, Google is assembling a massive knowledge base. It turns out that cars get better at driving in the same way that humans do, by logging a lot of miles. With its approach, Google is in position to learn faster than anyone else.

By contrast, even though some might have comparable technology capabilities, car makers are limited in their learning by their lower aspirations. One Daimler engineer captured this sentiment well in a recent interview:

City traffic is an utterly chaotic situation, and designing autonomous cars that can drive in it is not even one of our goals at this point. Autonomous driving in monotonous, steady highway traffic is a far more reasonable and feasible goal.

Similarly, by making critical assumptions about human drivers wanting or needing to remain in the loop, car makers limit their explorations and, thus, limit their ability to respond strategically if their assumptions turn out to be wrong.

More generally, by not having robust learning mechanisms in place—perhaps as just a hedge against implausible disruptive scenarios—car makers risk being caught flat-footed if the future takes an unexpected turn.

Our research shows that companies often fail because they have neither the time nor the inclination to learn, and end up swinging from thinking small to betting big. Betting big after falling behind is a recipe for failure. In the research for our previous book, “Billion-Dollar Lessons,” Paul Carroll and I determined that fully 46 percent of the 2,500 failures we investigated never had a chance to succeed, no matter how good the implementation, because the companies didn’t take the time or apply the discipline necessary to get the strategy right before beginning the rollout.

But how will Google profit?

Many readers of the first three parts of this series have said that I’m unrealistic, that Google knows nothing about cars and would never make them. Even if driverless technology is valuable, the reasoning goes, Google will not be a major player. Forbes contributor Haydn Shaughnessy left no doubt where he stands by following up his “nonsense” article with this:

Last week I said the idea of a Google driverless car is nonsense and I repeat it—Google will not be a force in autos.

But Google doesn’t have to supplant the car makers to be a formidable player, or to win. There are plenty of opportunities short of putting the car companies out of business.

If the driverless car technology works, Google could provide the driverless software to manufacturers, rather than making cars itself. That operating-system approach worked great for Microsoft, while the companies that actually made PCs struggled to make any kind of profit. And, of course, Google has already pulled off the strategy with its Android mobile operating system, which now leads Apple’s iPhone in market share.

More than 95 million new passenger and commercial vehicles were sold worldwide in 2012. All of these vehicles already include thousands of dollars of standard and optional safety features.

Let’s assume, to make the math simple, that a robust driverless system is worth somewhere between $500 (basic curtain air bags) and $5,000 (rough total of premium safety features on chart). That would equate to a market of $47.5B to $475B annually, not including any additional information services. Clearly, the market demand for ever more safety could easily support a robust, proven driverless capability.

Major car makers would, of course, resist letting Google control the driving OS layer of their vehicles. But, even if some carmakers can match Google’s technology, it is doubtful that every carmaker will succeed. So it is possible that some will turn to Google as a white knight in response to capabilities developed by GM, Daimler or Toyota.

The value of supplying driverless software to some portion of all new cars might warrant Google’s pursuing the Android strategy again by licensing the software for a price so low that manufacturers cannot refuse. Imagine a fleet of millions of cars feeding map and traffic data to Google maps, feeding location and behavior data to Google’s customer intelligence, acting as repeaters to Google’s broadband mesh WiFi network and, of course, exchanging queries and advertising via Google’s search engine.

How much are these opportunities worth to Google? The short answer is: a lot, potentially—especially since many enhance Google’s immensely profitable information-based businesses, rather than just threathen margin-poor automotive realms. But resist the urge to bring out your sharp pencils and smother the program with traditional CFO-type financial analysis.

Yes, Google will have to develop greater clarity about its commercialization strategies at some point. Right now, however, it has carved a competitive position with tremendous option value. It should continue developing its capabilities while keeping its strategic options open.

Remember the Law of Disruption that I discussed in Part 3. Google is, essentially, pursuing the exponentially advancing technology curve while car makers are betting on the incremental curve.

The key to Google’s success will be to accelerate the maturation of its technology-driven, automation-focused strategy while other players are focused on incremental approaches. Can it demonstrate that its holistic approach works while the automotive industry is still advocating partial solutions? If so, it will unleash market and regulatory pressures to adopt its solutions—because a lot of lives and dollars will be at stake. In effect, Google could create a demand that it is best positioned to fulfill.

Google’s Job One

As I discussed in Part 3 of this series, on adoption scenarios, Google is well positioned to jumpstart the adoption process with large-scale demonstrations and pilots. This would continue to develop the Google cars’ driving capabilities and enhance the company’s learning lead.

Following recent legislative victories in California, Google could deploy hundreds of Google cars to drive Googlers and others around the state. This could quickly generate millions of road miles and learning opportunities. It would accelerate efforts to crack lingering technology problems. It would create numerous opportunities to engage regulators and other influencers. And, it would generate mountains of evidence on the safety and benefits of the car.

Google could then move to pilot the technology at a larger scale. It could, perhaps, do this in Las Vegas, because Nevada has also approved the car, or in Kansas City, where the infrastructure that the company is laying for its Google Fiber rollout could be leveraged. Google could use its deep pockets to invest in supporting infrastructure (like road sensors). It could take the liability issues off the table by essentially self-insuring. And it could use its deep pockets to make the cars available at competitive prices.

Google could tackle this pilot on its own, or it could be a partner in the Big Venture Play scenario that I offered in Part 3 of this series. Either way, a regional demonstration would mirror the Google approach in Kansas City to demonstrate the viability of high-speed fiber networks to the home.

To gain further footholds, Google could target narrow, high-profile applications to gain experience and build support. It could, for example, focus on large trucks. Large trucks were involved in accidents that accounted for almost 10% of road fatalities in the U.S. in 2009, and truck drivers are five times more likely to die in work-related accidents than the average US worker. A large trucking pilot would also begin to illuminate what some consider a massive disruptive opportunity: logistics and supply chain reimagination enabled by driverless transportation.

Google could also start to explore alliances; if, as CEO Larry Page predicts, the driverless car will be production-ready in 5 to 10 years, it’s not too early to prime the pump for business opportunities, given how radical the change will be.

Google might work with upstarts in rapidly growing markets, perhaps a Tata Motors in India or Volvo Cars, which has long embraced safety as its core brand proposition and is now owned by Zhejiang Geely Holding Group of China. (A recent study estimated that 66 million Chinese households will be able to buy a new car by 2013 — a level 50 percent above the number of cars currently on the road in China.)

Imagine if a Google alliance with Tata or Volvo somehow generated the modern-day, driverless equivalent of the Volkswagen Beetle and captured the fancy of India’s or China’s emerging middle class? The alliance would then be positioned to sweep through other markets.

Once proven, there remains what some regard as the unimaginable option: Google could make its own cars. Google is unlikely to start its own car company, but it could certainly buy one. Remember that Geely bought Volvo for only $1.8 billion in 2010. Would a price tag even ten times higher be unimaginable given Google’s purchase of Motorola for $12.5 billion?

* * *

The complicated business-model considerations for driverless cars have already kicked off an intense game of three-dimensional chess, and Google could obviously lose. But, if it continues to Think Big, Start Small and Learn Fast, Google will accelerate the development of driverless capabilities worth trillions to society. It should also generate truckloads of profits for Google.