It is generally agreed that most markets are demand-driven. Supply will go where there is a demand to be met.

This is not always true, as the last few years have shown.

I had a conversation with someone who mentioned that while Lyft was working on innovative features for riders, Uber’s focus in the US was on supply. In Bengaluru, a lot of the demand has shifted to Namma Yatri and Rapido, because that is where the supply is today.

With the advent of self-driving cars, the platform power shift will again be in the direction of who controls the demand. If self-driving cars are the future, and everyone has one, then most of the people would rent out their car during the times when they are not using it. And they would be renting it out to the platform that has the most demand for it.

Caveat: The demand in the city where self-driving works. Demand isn’t global. It’s local. Ridehailing has local network effect, not global like in social networks.

There is an amazing post by Kevin Kwok about underutilised assets, and how most new age on-demand startups grew by exploiting them. You have an unused room in your house. You put it on Airbnb, but someone has to look after the guest. There is manual work involved. You have a car that you don’t use much. Y can drive it in your free time as a gig worker and earn some extra income.

In both of these cases, there are two components: the unused asset (the extra room, the car) and the labour (you).

Labour is not free.

Once self-driving becomes mainstream, there will be no more drivers. No labour involved in ridehailing. Waymo has already accomplished it in SF.

Fast forward to 2030. Let’s say a lot of manufacturers have cracked self driving by then. At least for select cities. Aggressive timeline for sure. But again, let’s assume it happens.

You are a software engineer in SF. You have a self-driving car. You are going to rent out your car. Why wouldn’t you? Your are decoupled from the fixed asset you own. You don’t have to drive the car the way people do today if you want to take advantage of the on-demand ridehailing platforms.

Another aspect of ridehailing that very few people talk about: Ridehailing loyalty programmes are very hard to crack.

A loyalty programme offers either monetary or non-monetary benefits. With Singapore Airlines Krisflyer Gold, you get to stand at the front of the queue for boarding. Marriott offers free room upgrades.

In ride-hailing, you can definitely create tiers of users based on a loyalty programme. But the only non-monetary benefit you can probably give is priority. But a lot of ride-hailing apps already allow users to raise their bid by paying more (whether it’s an advanced tip or something else), and if the supply is low, then you won’t see the benefit anyway.

You can’t upgrade the user from a standard car to a premium car in the same way that airlines can upgrade you from an economy seat to a premium seat. If there is a premium seat that is not being used, then there is no additional cost to the airline. Same goes for upgrading with Marriott.

There is no under-utilised/ unused asset in ridehailing that you can give away for free or at a discount. Someone has to pay for it.

With self-driving, unused cars are going to be a lot like empty seats in business class. Or the executive room at the Marriott for that matter.

You could even give your car away at a heavily subsidised rate if it was going to sit idle anyway. Uber, who will have all these cars on hire, will have a lot more flexibility in terms of how much they can charge. Yes, it is not a free, unused asset like the unoccupied room, but if you are making X dollars on your idle car, you are probably going to be OK with that.

What this means is that the platform that owns the majority of the demand will be able to do a lot more with its loyalty programme as well as its subscription programme than it can do today.

That is why I am bullish on Uber. It owns the most demand worldwide today. And self driving will make it even more dominant.