Today’s post comes from David Moss, a former LiDAR salesman and independent AV tester who has extensively tested autonomous vehicles from companies including Tesla, Waymo, and Zoox.
I watched this previously on David's YT channel. Wanted to check in here. Lyft Nashville is an interesting market. Nashville is the current ragged edge for Autonomy due to the VERY modest population density coupled with a VERY LOW population density compared to everywhere else Waymo has explored. Most places that have become launch cities for autonomous services have the 'right density' and a reasonable deadhead range I would think a tight partnership with Lyft improves the economics greatly in this case. I have that only knowing the urban core/metro population, the core density and the deadhead range is a remarkable predictor of where Waymo will go next.
I cut Lyft a bit of slack about someone ordering and canceling 14 rides in a row. A reminder of Uber's campaign in 2014, when 177 Uber employees deliberately ordered and canceled 5,560 fake rides over a 10-month period.
Your situation was not what Lyft's agreement with Waymo was intended to do, which is to provide a win-win outcome for you, Waymo and Lyft by providing one-stop access to a combined vehicle fleet.
I'm not surprised that early in this integration, Lyft driver supply may outweigh Waymo availability in a particular neighborhood and time of day. Of course, if you really were determined to take a Waymo ride, you would have ordered one on their app. Instead, you created an annoying inconvenience for Nashville Lyft drivers.
I agree Lyft’s behavior here was reasonable, but it was also a good stress test for the system. I could see other users doing the same thing if they really want a Waymo through Lyft and don’t realize they can just request one in the Waymo app. I can’t imagine Lyft is going out of its way to advertise that :)
And the drivers did get a $2 cancellation fee, so even if each cancellation caused a two-minute disruption, that works out to $60/hour, which as you know is a lot more than the average Lyft driver typically makes :)
While I fully understand the goal as a journalist to write about the experience of getting Waymo via Lyft, deliberately trying for it is a silly thing and it's not very surprising that it ended up with unwanted results.
Waymo's partnerships with Lyft and Uber are there to let companies like Lyft experiment with having robotaxis, and for Waymo to get extra ride requests when they want them. If Waymo is already getting enough requests with their own app, there's not much reason for them to get Lyft ones.
I think it was actually helpful for Lyft since I'm guessing many Lyft users have no idea that you can also get a Waymo directly on the Waymo app. So if you take a Waymo once on Lyft, enjoy it and then want to ensure you get one again, this would be normal behavior. Might be a good idea for Lyft to even add an 'AV only' option.
Excellent article — precise, and clearly the result of thorough work.
I have a few questions:
* Do you think Lyft prioritizes a human driver because it earns more from that ride, or simply because no Waymo vehicles were available?
* Did you try ordering a Waymo directly at the same time? Do you think Waymo prioritizes direct bookings through its own app over Lyft bookings, since it does not have to pay Lyft a commission?
* Could Lyft be deliberately avoiding an AV-only button?
* In what ways is a ride in a Tesla superior to a Waymo ride?
I watched this previously on David's YT channel. Wanted to check in here. Lyft Nashville is an interesting market. Nashville is the current ragged edge for Autonomy due to the VERY modest population density coupled with a VERY LOW population density compared to everywhere else Waymo has explored. Most places that have become launch cities for autonomous services have the 'right density' and a reasonable deadhead range I would think a tight partnership with Lyft improves the economics greatly in this case. I have that only knowing the urban core/metro population, the core density and the deadhead range is a remarkable predictor of where Waymo will go next.
I cut Lyft a bit of slack about someone ordering and canceling 14 rides in a row. A reminder of Uber's campaign in 2014, when 177 Uber employees deliberately ordered and canceled 5,560 fake rides over a 10-month period.
Your situation was not what Lyft's agreement with Waymo was intended to do, which is to provide a win-win outcome for you, Waymo and Lyft by providing one-stop access to a combined vehicle fleet.
I'm not surprised that early in this integration, Lyft driver supply may outweigh Waymo availability in a particular neighborhood and time of day. Of course, if you really were determined to take a Waymo ride, you would have ordered one on their app. Instead, you created an annoying inconvenience for Nashville Lyft drivers.
I agree Lyft’s behavior here was reasonable, but it was also a good stress test for the system. I could see other users doing the same thing if they really want a Waymo through Lyft and don’t realize they can just request one in the Waymo app. I can’t imagine Lyft is going out of its way to advertise that :)
And the drivers did get a $2 cancellation fee, so even if each cancellation caused a two-minute disruption, that works out to $60/hour, which as you know is a lot more than the average Lyft driver typically makes :)
While I fully understand the goal as a journalist to write about the experience of getting Waymo via Lyft, deliberately trying for it is a silly thing and it's not very surprising that it ended up with unwanted results.
Waymo's partnerships with Lyft and Uber are there to let companies like Lyft experiment with having robotaxis, and for Waymo to get extra ride requests when they want them. If Waymo is already getting enough requests with their own app, there's not much reason for them to get Lyft ones.
I think it was actually helpful for Lyft since I'm guessing many Lyft users have no idea that you can also get a Waymo directly on the Waymo app. So if you take a Waymo once on Lyft, enjoy it and then want to ensure you get one again, this would be normal behavior. Might be a good idea for Lyft to even add an 'AV only' option.
Excellent article — precise, and clearly the result of thorough work.
I have a few questions:
* Do you think Lyft prioritizes a human driver because it earns more from that ride, or simply because no Waymo vehicles were available?
* Did you try ordering a Waymo directly at the same time? Do you think Waymo prioritizes direct bookings through its own app over Lyft bookings, since it does not have to pay Lyft a commission?
* Could Lyft be deliberately avoiding an AV-only button?
* In what ways is a ride in a Tesla superior to a Waymo ride?