Ted Goranson - Personal Blog

I think this is the path they have committed to. The convention now is that every family owns a car or two so that they are available. This drives a great many dimensions in life. It enables suburbs and commuting, which in turn drags retail, education and city services behind it. It also feeds a market for car purchases and maintenance that is a huge industry. Among non-government enterprises, I think it ranks only behind health care.

What if we did not need to own a car; what if the availability of an Uber car was close enough to having one in the driveway? To enable this, Uber has to get into self-driving AI, self-organizing distribution control and car manufacturing. They would also get into errand support, like delivery of on-line purchases from local outlets.

This is why the market is excited. That business of building cars to sit more than 90% of the time is ripe for disruption. If I were in charge, I wouldn’t worry too much about the self-driving AI. That’s nearly here. I would worry a bit about car manufacturing, but they don’t need to revolutionize that in order to co-opt it. (It is revolutionizable.)

I’d worry instead about how vehicles self-distribute. The core notion behind Uber is that it is self-organizing. They provide a way for customers to see cars, and drivers to see customers. They depend on decentralized decisionmaking for when a driver comes on line, where he/she decides to be and (to some extent) whether the driver will bid on the job. The whole business model depends on drivers being conveniently available.

We know a lot about influencing these self-organizing systems. The prevailing approaches won’t work. Were I in charge, I would invest in this. If you own this, you own a lever for disrupting many things at the edge.

Were I in charge, a business I would look closely at is long distance packages. This is a huge business. Amazon and Google think the way out is drones, but this won’t scale. A key risk is theft.

The need for package delivery will increase as internet retail commerce increases. The two models we have are at their limit.

Big Box Delivery

The smaller of those two models is big box distribution. In this model, you order from Walmart on line. Much of their on line catalog is also available in the store. Basically, someone takes it off the shelf in a location near the consumer and either arranges a pickup or does some local delivery from store to consumer.

The advantage of this for Walmart is that the one thing they do better than anyone is keep those stores stocked at low cost. Walmart is a goods distribution company and a retailer incidentally. There are still hub and spoke geometries behind the scenes but most of that is pushed onto the wholesaler.

As it happens, no big box retailer can crack this. They need to maintain their stores, and it makes great sense to leverage the fact that they are everywhere and that they have smooth lines all the way back to Chinese manufacturers. Simply adding Uber delivery for the last mile is not enough.

Hubs

The other model is maintaining genuine hubs. Some of Amazon’s business has suppliers managing stock and shipping. An order comes in for an item, it is forwarded to Otherguy Inc who prints an Amazon receipt and ships an Amazon box. This model is likely to grow, but not as a percentage of Amazon’s business. It still makes sense for Amazon to build huge semiautomated warehouses that have vast amounts of stock. The packages still travel hundreds of miles because the courier services (UPS/USPS, and Fedex) primarily use hub and spoke.

They have to because of the management of capital in vehicles and the scheduling of drivers/pilots. This system is disruptable.

What if there were millions of drivers, each able to be part of a virtual, temporary pony express to get goods from hundreds of thousands of suppliers (with no wholesaler) to hundreds of millions of consumers? A box could go through dozens of vehicles and if from Asia go through a consolidator.

The numbers work out because each contractor is carrying the capital burden of the vehicle and the costs of down time. All you need is a way for packages to bid on transport the same way that humans do. A complication is that you want to bid on collaborative transit streams; the package will only be at such and such parking lot for you if a group of upstream conveyers commit, otherwise it will be at another handoff location.

This is not as difficult a problem as the self-organizing distribution problem, at least not at first. If you crack it, you will own more than half of what Amazon wants to own.

Advertising

This is a subject for another post, but clearly the agent that delivers the goods can be the agent that orders the goods. And that agent is one that can move up the food chain to what we now call advertising and brand management. Advertising is profoundly broken. Google and Facebook will break it more. I think advertising will bifurcate into groups who spy on you and push ads in exchange for something desirable.

The other path is that when discovery is reinvented, replacing search, goods and services will a primary source of information. This also may have some benefits attached, but be on the pull side of the equation. I can see how this would work.