How the Robot Taxi Supply-as-a-Service Model Will Be Executed (Part 3 of 3)
How four critical areas of the supply-as-a-service model will be structured between suppliers of autonomous vehicles and the ride-hailing platforms.
Part 1 of this three-part series explains why the SaaS model will be so critical to robot taxi commercialization. Part 2 provides a framework and detailed considerations to guide the execution of the model. In this final Part 3, I apply the framework and considerations from Part 2 to predict how the model will be executed along four critical and complex dimensions of the model.
Vehicle control policies between the platform and AV supplier will solve problems in various categories, namely: A) Fleet utilization, B) Vehicle maintenance, C) Rider experience, D) vehicle routing, E) Remote pilot operations.
Each AV supplier will provide each platform with a simple geo-fence that specifies where in each metro area the vehicle can pick up and drop off riders.
AV supplier owns vehicle routing and hence plays significant role in vehicle dispatching. Before a ride is dispatched to the AV, the rider must be quoted a fare, which is computed by the platform after making a call to AV supplier’s routing system. The AV supplier then shares the rider’s booked-route back to the platform for potentially adding passengers to a ride-chain (e.g. UberPool, Lyft Line).
The platform will control inter-ride vehicle positioning during the rental period, during which the platform also has exclusive access to the vehicle.
During the rental period, both the platform and AV supplier will have the ability to take a vehicle off-line for safety or maintenance measures.
At the end of the rental period, vehicle is returned to the custody of the AV supplier.
Pricing and revenue model will be a hybrid of fixed, time-based fee plus miles-based ride fee. Platforms have a natural incentive to maximize supply, demand fulfillment, and minimize surge pricing because this protects its most valuable asset: the rider experience and brand. Hence an AV supplier would only need to worry about the platform purposefully underutilizing it’s vehicles if a competing AV supplier offers lower pricing. Since it is highly likely that platforms will use competing AV supplier’s vehicles simultaneously on the platform, per-ride pricing would cause platforms to send demand to the lowest-cost provider. This would create uncertainty of fleet utilization for the AV supplier.
As a result, to align the platform’s interests with the AV supplier’s, I expect the pricing/revenue model to be a two-part model consisting of time-based rental fee and per ride fee. The platform will agree to pay in-advance a given price for a given unit of vehicle capacity (measured in vehicle hours) at a specific time and location (e.g. 10 vehicle hours, Monday, 2–4pm, within geo fence defined by set of GPS coordinates).
The base rental fee acts as a minimum utilization guarantee to cover the AV supplier’s time-based depreciation and opportunity costs. The per-ride price that is a function of both ride length (miles) and duration (time), covers the AV suppliers miles-based depreciation of the vehicle. Platform will cover the cost of miles driven between rides.
Platforms will require a SLA to ensure the supplier provides agreed-upon supply when and where they specified.
Two things the AV supplier will not have direct control over:
- the platform’s demand-fulfillment strategy (i.e. using AV supply to fulfill ‘base’ or ‘peak demand’)
- whether the platform shares surge pricing revenues with it.
Data sharing will be intentionally limited. AV suppliers will not get all the data they want and yet, they will get more than the platform would like them to have. Suppliers will be able to capture some ride-demand data as the rides are dispatched and completed by their vehicles. Suppliers will however lack complete visibility into the entire market demand curve because not all rides will be fulfilled by their cars. Platforms will not share PII of their riders.
Branding will be a battle that AV suppliers win. Today, riders care little about the driver’s vehicle brand, and the platforms want it this way. In a robot-taxi future, however, riders will care much more about the vehicle brand because the vehicle — not the driver — dictates so much of the rider experience. Riders will be trusting the vehicle brand instead the platform’s ability to screen safe drivers. Riders will grow to develop trust and preference for vehicle build and maintenance quality, cleanliness, in-vehicle experiences, and overall vehicle design and interaction experience. This will be especially true during early commercialization phase, when early-adopter riders will be the primary source of demand. As a result, platforms will have to cater to the rider’s need to know the brand of the AV, and this will provide the AV supplier with brand-equity-building opportunities in the rider app, platform communications, and the vehicle’s interior and exterior.
Final Thoughts
These next 3–5 years of robot taxi commercialization will be hugely influential to the longer-term future of robot taxis and the industry ecosystem/structure; and the Supply-as-a-Service model will play a massive role in these next 3–5 years of commercialization. Interestingly, I predict however that the SaaS model will affect the longer-term horizon of the robot taxi industry only in so far as it impacts the hugely foundational near-term 3–5 year period. Beyond 5 years, I predict the fundamental network effects underlying the robot taxi industry will drive vertical integration and consolidation amongst competitors, yielding the SaaS model all but irrelevant.
Part 2 — Executing A Robot Taxi Supply-as-a-Service Model
Part 1 — The Role of the New SaaS Model in Our Robot Taxi Future
