As someone in Los Angeles who uses Waymo at least weekly, my experience is that pricing has become a bigger factor over the past few months. Rides that often used to cost me around $13-$17 before promotions are now regularly over $20. I enjoy using the service and would like to use it more, but at those prices I often choose to drive myself unless parking is difficult or I'm planning to drink.
If prices came back down, I could easily see myself using Waymo most days, potentially even for my daily commute. That makes me wonder how much pricing is affecting demand and ridership growth.
Coverage is another limitation. I would use Waymo more if it served destinations such as Woodland Hills, Pasadena, and LAX. If either my starting point or destination is outside the service area, the trip simply isn't an option.
The pickup and drop-off experience can also be frustrating at times. Waymo doesn't always allow the exact front entrance of a destination to be used, so pickups and drop-offs can end up down the street or even several blocks away. It's understandable that there are operational and safety reasons for some of these decisions, but it does add friction compared to driving yourself directly to the door.
Overall, I still like the service, but higher prices, service-area limitations, and pickup/drop-off constraints are the main reasons I don't use it more frequently.
Excellent point about pricing possibly shaping the demand curve and where supply/demand equilibrium falls. I also think the distance per trip piece is important because the per mile revenue profile looks different for a 10 trip compared to a 2 mile one. We don’t have public pricing data to my knowledge.
I live in Detroit, was in SF a few weeks ago on a family trip, and ordered my first Waymo for my mobility-challenged mother and I to travel 4 miles to our bnb. Both the pickup and the drop off were hundreds of yards from what I called for, and became quite a challenge for her. (Notably, my receipt states the intersections I called for, not the actual PUDO locations.)
I wrote a note on the experience to LinkedIn, making 46,000 impressions 😳, plus 40 comments and was mentioned in a Center for Auto Safety podcast. The story touched a nerve.
I teach a graduate Systems Thinking course to Detroit-based auto engineering managers. We cover some innovation rules:
Customers will make their purchasing decisions based on stepwise criteria:
1. Functionality
2. Reliability
3. Convenience
4. Price.
From the providers perspective
1. What are the Jobs to be Done?
2. What assumptions must hold true to accomplish these?
3. What are customer’s alternatives?
Last, let’s look at diffusion of innovation and market segmentation
1. Early Adopters
2. Visionaries
3. Early Majority (where the money is)
4. Late Majority (even more money)
5. Laggards.
In each of these dimensions, the Robotaxi business model seems weak, and if they can’t reach recurring use (killer app status) amongst the Early Majority, this is not a plateau so much as a peak.
interesting, thanks for sharing. I do think the pick up and drop off experience is going to be a big challenge for all autonomous vehicles since you essentially have to program the vehicle to constantly break the law.
My next article (and I’m using your articles and those of Matthew Raifman in my research) is tentatively titled “Waymo v. Erlang”, and contrasts the current AV business model to the queueing theory developed a century ago when scaling telephone service in large cities. A key difference is that idle telephone lines do not consume resources from the public good - unlike AVs (streets and parking). Hence, Waymo’s circle through neighborhoods, ticking off the residents and generating NIMBYism.
PUDO doesn’t necessitate breaking the law. Busses and trains do it all the time. Delivery vans double park, but a simpler autonomous solution than AVs would be little robots ejected from the truck - going to the door as it circles. Or dedicated package delivery devices dropping off at a nearby lockbox during the night.
Interested in your feedback, whenever I get this thing done. 🙄
Interesting post! One clarification regarding the “waiting time” — I believe this actually refers to vehicle waiting time rather than passenger waiting time. So excessive vehicles on the road relative to demand would cause this number to increase and vice-versa. Here’s what found in the glossary for the “TotalWaiting” field:
“The total amount of time vehicles waited between ending one passenger trip and initiating the next passenger trip, expressed as a monthly total in hours
Waiting begins after end of previous trip's Period 3 (passenger drop off) to beginning of next trip's Period 2 (request accepted, vehicle en route to next passenger).“
My interpretation of the stable pattern since summer 2025 is that Waymo has reached an “optimal” policy with respect to dispatching cars into service vs leaving them at the depot so as not to accrue excessive deadheading mileage.
Yes that is correct, it's the 'waiting time' of the vehicle or essentially deadheading time (Period 1). The lower this number the better. Period 2 is a similar metric in that you want it to be as low as possible since that is also essentially unpaid time and the lower the number, the higher the overall efficiency.
Great topic. I can theorize a lot of reasons why the miles might flatten out. It will be interesting to see the statistics for the safety report thru the end of Q1 soon for Waymo (maybe two weeks from now). They were covering about 160,000 autonomous miles daily in the Bay Area in Q4 2025. and about 134,000 daily in Los Angeles. Matthew :: are you forecasting those miles will flatten or fall in the Waymo Safety Report?
I think they might fall but it’s something we could calculate from the CPUC data. It includes total revenue miles by quarter. Waymo is doing fewer rides but they are longer, so I would expect any fall off for miles traveled to look slightly attenuated compared to the trip curve.
As someone in Los Angeles who uses Waymo at least weekly, my experience is that pricing has become a bigger factor over the past few months. Rides that often used to cost me around $13-$17 before promotions are now regularly over $20. I enjoy using the service and would like to use it more, but at those prices I often choose to drive myself unless parking is difficult or I'm planning to drink.
If prices came back down, I could easily see myself using Waymo most days, potentially even for my daily commute. That makes me wonder how much pricing is affecting demand and ridership growth.
Coverage is another limitation. I would use Waymo more if it served destinations such as Woodland Hills, Pasadena, and LAX. If either my starting point or destination is outside the service area, the trip simply isn't an option.
The pickup and drop-off experience can also be frustrating at times. Waymo doesn't always allow the exact front entrance of a destination to be used, so pickups and drop-offs can end up down the street or even several blocks away. It's understandable that there are operational and safety reasons for some of these decisions, but it does add friction compared to driving yourself directly to the door.
Overall, I still like the service, but higher prices, service-area limitations, and pickup/drop-off constraints are the main reasons I don't use it more frequently.
Excellent point about pricing possibly shaping the demand curve and where supply/demand equilibrium falls. I also think the distance per trip piece is important because the per mile revenue profile looks different for a 10 trip compared to a 2 mile one. We don’t have public pricing data to my knowledge.
Waymo is definitely is more expensive during surge times since the only lever they have to temper demand is raising prices - https://x.com/TheRideshareGuy/status/2055842619482423558?s=20
I’m intrigued by your PUDO comments.
I live in Detroit, was in SF a few weeks ago on a family trip, and ordered my first Waymo for my mobility-challenged mother and I to travel 4 miles to our bnb. Both the pickup and the drop off were hundreds of yards from what I called for, and became quite a challenge for her. (Notably, my receipt states the intersections I called for, not the actual PUDO locations.)
I wrote a note on the experience to LinkedIn, making 46,000 impressions 😳, plus 40 comments and was mentioned in a Center for Auto Safety podcast. The story touched a nerve.
I teach a graduate Systems Thinking course to Detroit-based auto engineering managers. We cover some innovation rules:
Customers will make their purchasing decisions based on stepwise criteria:
1. Functionality
2. Reliability
3. Convenience
4. Price.
From the providers perspective
1. What are the Jobs to be Done?
2. What assumptions must hold true to accomplish these?
3. What are customer’s alternatives?
Last, let’s look at diffusion of innovation and market segmentation
1. Early Adopters
2. Visionaries
3. Early Majority (where the money is)
4. Late Majority (even more money)
5. Laggards.
In each of these dimensions, the Robotaxi business model seems weak, and if they can’t reach recurring use (killer app status) amongst the Early Majority, this is not a plateau so much as a peak.
https://www.linkedin.com/feed/update/urn:li:activity:7489504416257966080?updateEntityUrn=urn%3Ali%3Afs_updateV2%3A%28urn%3Ali%3Aactivity%3A7489504416257966080%2CFEED_DETAIL%2CEMPTY%2CDEFAULT%2Cfalse%29
interesting, thanks for sharing. I do think the pick up and drop off experience is going to be a big challenge for all autonomous vehicles since you essentially have to program the vehicle to constantly break the law.
Thanks for following me!
My next article (and I’m using your articles and those of Matthew Raifman in my research) is tentatively titled “Waymo v. Erlang”, and contrasts the current AV business model to the queueing theory developed a century ago when scaling telephone service in large cities. A key difference is that idle telephone lines do not consume resources from the public good - unlike AVs (streets and parking). Hence, Waymo’s circle through neighborhoods, ticking off the residents and generating NIMBYism.
PUDO doesn’t necessitate breaking the law. Busses and trains do it all the time. Delivery vans double park, but a simpler autonomous solution than AVs would be little robots ejected from the truck - going to the door as it circles. Or dedicated package delivery devices dropping off at a nearby lockbox during the night.
Interested in your feedback, whenever I get this thing done. 🙄
sure, send me an email with details and I’m happy to review the article and provide feedback before you publish. Harry@therideshareguy.com
Excellent post, Matt! Is it feasible to set up a self-updating database like Todd Schneider does for NYC ridehailing data: https://toddwschneider.com/dashboards/nyc-taxi-ridehailing-uber-lyft-data/
Love this idea. Let’s work on it!
Interesting post! One clarification regarding the “waiting time” — I believe this actually refers to vehicle waiting time rather than passenger waiting time. So excessive vehicles on the road relative to demand would cause this number to increase and vice-versa. Here’s what found in the glossary for the “TotalWaiting” field:
“The total amount of time vehicles waited between ending one passenger trip and initiating the next passenger trip, expressed as a monthly total in hours
Waiting begins after end of previous trip's Period 3 (passenger drop off) to beginning of next trip's Period 2 (request accepted, vehicle en route to next passenger).“
My interpretation of the stable pattern since summer 2025 is that Waymo has reached an “optimal” policy with respect to dispatching cars into service vs leaving them at the depot so as not to accrue excessive deadheading mileage.
Yes that is correct, it's the 'waiting time' of the vehicle or essentially deadheading time (Period 1). The lower this number the better. Period 2 is a similar metric in that you want it to be as low as possible since that is also essentially unpaid time and the lower the number, the higher the overall efficiency.
Illuminating. Thank you 🙏
Great topic. I can theorize a lot of reasons why the miles might flatten out. It will be interesting to see the statistics for the safety report thru the end of Q1 soon for Waymo (maybe two weeks from now). They were covering about 160,000 autonomous miles daily in the Bay Area in Q4 2025. and about 134,000 daily in Los Angeles. Matthew :: are you forecasting those miles will flatten or fall in the Waymo Safety Report?
I think they might fall but it’s something we could calculate from the CPUC data. It includes total revenue miles by quarter. Waymo is doing fewer rides but they are longer, so I would expect any fall off for miles traveled to look slightly attenuated compared to the trip curve.
Thanks for taking the time and a well fashioned answer.