In the fall of 2002, Microsoft announced the Smart Personal Objects Initiative (SPOT), which aimed to inject software into everyday objects like watches and pens to make them smarter. It projected a very large target market for SPOT devices, with keychains, pens and watches alone potentially amounting to “hundreds of millions of units, even billions of units a year.” In reality, SPOT was a miserable commercial failure.
Fast forward to 2014 when Apple introduced the Apple Watch with great fanfare as a fashion and lifestyle product. It projected sales of 40 million units in the first year, but actual sales for the Apple Watch were less than 10 million units. It wasn’t until Apple repositioned the watch in its fourth version as a more focused health and fitness device that sales improved.
Mark Twain observed, “History doesn’t repeat itself. But it sure does rhyme!” I see a common pattern: companies that create new technology products tend to focus on the largest and most visible target market. They chase after the mass market, believing that “everybody is going to want one of these!” This is the Myth of the Mass Market, and following it is a huge mistake. For a new technology to appeal to the mass market, it needs to be optimized for a wide variety of customers and usage situations. It needs to be “everything to everybody,” but it usually ends up being “something to everybody” and “everything to nobody.”
Autonomous Vehicles Chasing Mythical Mass Market
The real short-term opportunities lie in less visible, more narrowly defined niche markets. This pattern is repeating itself in the autonomous mobility market. Automotive providers are desperately vying for a piece of the mass market for transporting people in cars. As I discussed in my last article, having driverless cars operating on any road, anywhere – the myth of substitution – is neither practical nor possible in the foreseeable future. The real market opportunity lies in niche markets and applications. While this may not look as sexy as the vision for driverless cars, it is a more pragmatic way to commercialize autonomous technology and develop it over time.
Pursuit of the mythical mass market for mobility is not without precedent. When Dean Kamien unveiled the Segway in 2001, he foresaw a transportation revolution and predicted selling 10,000 units a week by 2002. The real number was closer to 10 units a week. Consumers balked at paying thousands of dollars for a high-tech two-wheeled scooter that was too fast for the sidewalk and too slow for the road, when the alternative was to walk or take the bus. The mythical mass market never materialized for the Segway, but it did find two important niches. In public safety, police officers patrolling a beat and security personnel in warehouses or other large spaces and on campuses use the Segway Patroller to travel greater distances, maneuver in tight spaces, and not get fatigued. Tourism operators use the Segway successfully for city tours, allowing tourists to navigate crowded city streets and to see more than they could on foot.
Autonomous mobility will find promising areas of opportunity in various target markets. Autonomous mobility can be applied to transport people or cargo by land, sea or air. Between people and cargo, the more pragmatic opportunity is cargo. And among land, sea, and air, the best opportunities are sea transport and air transport. Even in land transportation, highways are more promising for autonomous mobility than urban transportation. Transporting people by land in urban settings turns out to be the most technically difficult and least economically feasible opportunity.
Cargo, not People
Transporting cargo is inherently easier than transporting people. Cargo doesn’t need to be protected at all costs. Cargo doesn’t need to be kept air-conditioned. Cargo doesn’t need to be transported to random locations based on the whims of passengers, and it doesn’t need to get to the destination as quickly as possible. In fact, humans are the most expensive, complex and demanding type of cargo. When it comes to choosing between autonomous mobility for cargo versus people, the answer is clear: go with cargo.
Autonomous cargo applications are developing at a rapid pace. Wilhelmsen of Norway, a large shipping company, and Kongsberg, a technology company, recently joined forces in autonomous ship design, development, logistics, and vessel operations. Walmart is testing delivery of goods from warehouse to warehouse on a fixed route. Volvo is testing an autonomous garbage truck for urban environments.
Sea and Air, Not Land
Autonomous mobility solutions work better on sea and in the air where there is less traffic, no pedestrians, and fewer complications from people and drivers behaving badly. Consider shipping on the high seas. The routes are repetitive and routine, so the journey from point A to point B can be programmed without worry of traffic or obstacles. Autonomous ships can dramatically reduce human errors, which are responsible for an astounding 96% of marine accidents. Autonomous ships can be much lighter because there is no need for crew quarters, galleys or lifeboats. Even the threat of piracy can be minimized without humans onboard.
In air transport, drones are being deployed for deliveries for many of the same reasons that we see in oceangoing transport. For example, Wing, an Alphabet subsidiary, was approved earlier this year for its first drone package deliveries in the United States, and UPS has applied to operate commercial delivery drones in the U.S. through its new UPS Flight Forward unit.
When you move to land, things get far more complicated because of hazards like jaywalking pedestrians, impaired drivers, road construction and emergency vehicles, to name a few. Autonomous vehicles must be programmed to deal with all possible “corner cases” such as encountering a pedestrian in the roadway on an icy road in heavy traffic at night.
Restricted Domains—Not Any Road, Anywhere
Even in ground transportation, it is easier for autonomous vehicles to handle restricted Operational Driving Domains (ODD) than to take any road anywhere. For instance, highway travel is easier than the first mile and the last mile of a road trip, because vehicles on a highway tend to move at similar speeds, and there are no traffic lights or pedestrians to deal with. Automation of highway driving mostly requires decisions on acceleration, braking, steering and lane changes – something Tesla’s cars have been able to do since October 2015.
Long-haul trucking will likely see the first widespread adoption of advanced driver assistance systems and full automation. Amazon, for example, is said to be using autonomous trucks developed by Embark to move cargo on the I-10 interstate highway. Convoys of driverless trucks have been tested in Europe as a way to address road safety and efficiency in fleet operations. Semi- and fully autonomous vehicles may also become more prevalent in restricted ODDs to transport both goods and people. These may include delivery trucks and autonomous vehicles in retirement communities, corporate campuses and golf courses.
While the mythical mass market calls seductively to autonomous vehicle developers, it’s a siren song that’s best avoided for now. Far more feasible and realistic are specialty uses of autonomous technology for moving cargo first and people later and on the sea and air first and land later. The lesson: if you chase the dead center of a market, you may be dead wrong.