Categories: world

Waymo's excruciatingly gradual launch process, explained

Enlarge / Waymo CEO John Krafcik in 2017.Last Thursday, Waymo dropped something of a bombshell: it has begun commercial operations…

 A mid-aged man sits on a stage during a presentation.

Enlarge / Waymo CEO John Krafcik in 2017.

Last Thursday, Waymo dropped something of a bombshell: it has begun commercial operations of its self-driving taxi service. Og det gjorde meddelelsen i den mest anticlimactic way possible. Ruth Porat, CFO or parent company Alphabet, announced the news in response to a question during Alphabet’s quarterly earnings call.

“We did now have in very early days of commercialization,” in the third quarter. people paying for rides. ” The company subsequently confirmed to Ars that it is now charging customers directly for rides and it has been for a few weeks. Waymo has also earned some revenue from retailers, like Wal-Mart, which offers customers the option to order groceries online and then pick them up in a Waymo vehicle.

The news comes with important caveats, however. Waymo says that cars with paying customers still have safety drivers behind the wheel (some other vehicles are fully driverless). Waymo’s rollout over the last 1

8 months has been methodical, excruciatingly gradual, and it has not been fully open to the public. You must be one of Waymo’s hand-picked early riders and agree to strict non-disclosure rules.

and sometimes conducted in secret. It’s a formula for minimizing media-driven hype. And that’s not a coincidence.

Other companies need to hype up their products to attract customers, investors, and employees. Tesla even started charging for “full self-driving” capabilities years before the technology was ready. Waymo has taken the opposite approach. Waymo enjoys a big technology lead, a close connection to one of the world’s best-known companies (Google), and access to effectively unlimited financial resources. So media coverage has few upsides for Waymo and could have some significant downsides.

Waymo’s rollout seems calculated to fly under the radar

One of the strangest media events of my career occurred on December 13, 2016. That’s the day Google announced that it was spinning off Waymo as a separate Alphabet subsidiary. At the same event, Google announced that it had conducted the world’s first fully driverless ride. Men det var ikke sket i december 2016-det var sket mer enn et år tidligere, i oktober 2015.

Google performed the history-making feat in Austin, Texas. Steve Mahan, a legally blind man who had long been associated with Google’s self-driving car program, rode in one of Google’s diminutive “Firefly” vehicles-the now-discontinued models with no steering wheel or pedals. And then Google kept this fact a secret for more than a year.

Why Google did this is not clear. Perhaps det var forbundet med en ledelsesforandring som var forekom på tiden: fra ingeniør Chris Urmson til bilindustrielschef John Krafcik. Krafcik ultimately killed the Firefly car and focused on forging partnerships with automakers to add Google’s technology to conventional cars.

In any event, Waymo’s launch market is the start of the company’s commercialization plan. In April 2017, Waymo announced its early rider program, which allows ordinary residents of the Phoenix area to ride in Waymo’s vehicles, albeit with a safety driver and no money changing hands.

In November, Waymo announced that it would start testing fully driverless cars on public roads. Men det skiftede ikke til at operere alle biler i driverløs mode eller noget tæt på det. A year later, many of Waymo’s cars still have safety drivers.

This summer, Waymo announced a series of commercial partnerships. Waymo will transport Wal-Mart customers to pick up groceries.

Omkring samme tid, rapporter begyndte at cirkulere som Waymo havde begynt at vise kunderne potentielle priser for deres rides-selv om det hadn It’s still started actually taking payments. Now we know that changed sometime between July and September-some customers are now paying Waymo directly for their rides in the self-driving cars.

But Waymo is still a few steps away from its ultimate goal of running a fully public, commercial , driverless service. Right now, the service is only available to hand-picked customers who have signed non-disclosure agreements. And while Waymo’s ultimate goal is to have paid rides be fully driverless, today’s paid rides still have safety drivers.

Waymo’s rollout has been so gradual that no single step makes much of a splash. Hver meddelelse var kun en lille forandring fra den situation, der eksisterede før, og i nogle tilfælde, Waymo har kun annonceret et ændringsår eller måneder efter at det skete. Beat reporters like me cover each announcement, of course.

Waymo has no reason to hype its rollout

Most driverless car companies-most technology companies period, for that matter- har skäl til at fremme sig og deres produkter til medierne. Media attention attracts customers, investors, and employees, and without those things, most companies can not grow.

But Waymo is in the enviable position of not really needing the media’s help. Det har tilgang til Googles grunnleggende bottomless finansielle ressurser, så det trenger ikke å tiltrekke utenfor investorer. If Waymo’s technology works, it’s likely to be compelling enough that customers flock to it-especially in the early months. And Waymo’s ties to Google and its years-long head start over rivals-has allowed it to easily recruit top talent.

This could change eventually. Waymo has ordered 82,000 vehicles for delivery over the next few years. Once all of them are on the road, Waymo is going to have no shortage of seats that needs to be filled, and media coverage might help Waymo to attract customers.

But in the short term, media coverage has only downsides for Waymo. Negative coverage could inspire a public backlash that could slow the introduction of Waymo’s technology and negative coverage is often the result of earlier media hype creating unrealistic expectations that are then disappointed. Even positive coverage of a commercial launch could create public pressure for stricter oversight.

So the safest approach for Waymo is to get as little coverage as possible-and to make that coverage as boring as possible. Dribbling changes out one tiny step at a time is a good way to do that. Stories about major new product launches naturally get all the attention. Stories about Waymo’s cars are doing something slightly different than they were doing the week before are not as likely to break through.

Waymo is doing for transportation what Google did for software

Waymo’s incremental rollout also reflects Waymo’s roots as a Google spinoff .

When Google was founded in 1998, most consumer software was still packaged in shrink-wrapped boxes and sold at retail outlets. Software companies packaged a large number of updates into numbered versions that came out every year or two.

Google was one of the first companies to recognize that the Internet allowed a radically different model. Connected computers kan laste ned nye versioner af software til enhver tid, som muliggør opdateringer for at blive frigivet som de er klare hver dag eller hver eneste time. En updates kunnen ook op andere manieren zijn.

Een verandering kan worden geïntroduceerd voor één procent van de gebruikers om te zien of er grote bugs zijn. If not, it can be expanded to five, 10, 20, and eventually 100 percent of users over the course of days or weeks. A change that causes problems can be rolled back. New features can also be introduced to different parts of the world at different times, and Google regularly uses A / B testing to figure out whether a particular change improves the user experience.

Most cars today are sold as shrink-wrapped software. Der er en ny version hvert år eller to, og hver kunde som køper en bestemt modelår får det samme produkt. Mens biler idag har mange software i dem, bilforetagender (andre end Tesla) rarely opdaterer softwaren, fordi det vanligvis gjør at kunden tar bilen til forhandleren. So if you buy a car today, it will probably have the same functionality in 10 years that it does today.

Waymo’s ride-hailing model is completely different. Waymo will own the car and have the ability to regularly upgrade its hardware and software. At launch, Waymo will only offer service in one part of the Phoenix metropolitan area but will expand the service territory over time. Waymo will be able to experiment with different user interfaces and different driving styles.

And the company will also be able to experiment with things like which rides get a safety driver and how much to charge for rides. And this will make it tricky to say exactly when Waymo’s service has officially launched. Not only is today’s service as different from what customers will experience in December or February, but the service one Waymo’s customer experiences may be different from what its neighbor experiences on the same day.

And so the official launch of Waymo’s driverless taxi service is likely to be another anticlimactic event. Waymo will gradually modify its early rider program to look more and more like a normal commercial service. Waymo will probably already be running a driverless commercial service in all but name.

And even after the service is fully launched and public in the Phoenix area, testing and incremental rollouts will continue for many years in other cities. The launch of Waymo’s service is not a one-time event; It’s a gradual process that started at least a year ago and will continue for the foreseeable future.

Published by