There are at least two macro trends* I think will unfold over the next 3-10 years, and put Uber on the leading edge of autonomous cars. I was only able to write up the first of them today.
TREND #1: There is a Storm Coming to the Car Mobility Value Chain
The standard model
To establish a frame of reference, let's simplify any given value chain into three parts:
***************************
Supplier - Distributor - Customer
Supplier - Distributor - Customer
Supplier - Distributor - Customer
***************************
So to build a great business, you either [1] get a horizontal monopoly in one part of the chain (i.e., own supply/own distribution/own the customer relationship); or [2] integrate two parts of the chain---usually within a vertical---for a differentiated value prop.
Before the Internet, a fundamental fact about the (Distributor - Customer) relationship was that
transactions are costly. Physical channels are expensive. To double the size of your channel, you will pay about double the fixed costs; and to double your transaction rate, it'll likely cost you about 2x the SG&A.
In a world where transactions are costly, it's super expensive for a distributor to win by owning the customer relationship at scale. You'll do better by integrating back into supply. The Internet turned that dynamic on its head, and I think mobile proliferation will do much the same, in way that has huge consequences for the auto value chain.
I'll start by highlighting the pattern in an illustrative example that contrasts Wal-Mart and Amazon.
Classical retail strategy
Wal-Mart is the canonical example of pre-Internet retail strategy. They invented vendor-managed inventory, and won suppliers by giving them unprecedented visibility into demand signals (while cutting costs in a million other ways.) Wal-Mart warped the value chain:
***************************
Vendor \ ____ / Shopper
Vendor - WMT- Shopper
Vendor / ____ \ Shopper
***************************
Of course, this was a Faustian bargain for the suppliers, since Wal-Mart now saw a strategic opportunity to own more customers; and that opportunity was the suppliers' margins.
Internet-era retail strategy
Speaking of Amazon. Think about how their strategy in books was fundamentally different---and impossible before the Internet. The bookselling value chain looked like:
***************************
Publisher - Chain - Customer
Publisher - Chain - Customer
Publisher - Indep - Customer
***************************
Amazon came out of right field to integrate directly with customers, at scale.
***************************
Publisher - Chain - Customer \
Publisher - Chain - Customer - AMZN
Publisher - Indep - Customer /
***************************
They COULD do this because on the Internet,
transactions are cheap. And they had a reason to do it, because they were creating a truly superior customer experience.
Convenience, selection, lower costs, yes; but also a shopping experience customized for you by machine intelligence, and social-proofed by the crowd. (Very importantly, both the last two are characterized by virtuous cycles. The more users they have, the better they become, the better they attract users...)
Better. Amazon shifted the battlefield to the user experience, where they had the permanent upper hand. And when you own the customers, the suppliers will come---upon which you can take a sledgehammer to their margins, in classical John Henry fashion.
Importantly, Amazon did this starting with zero physical assets. No distribution centers, no cutting-edge robots or logistic networks, nothing but the club of user experience and a violent predisposition.
Mobile-era car-for-hire strategy
We move next to a market the Internet
alone couldn't truly disrupt, because adding a channel in every living room wasn't good enough---you need a channel in every pocket/purse!
So, the traditional car-for-hire value chain was approximately:
***************************
Taxi Driver - Taxi Co. - Customer
Taxi Driver - Taxi Co. - Customer
Taxi Driver - Taxi Co. - Customer
***************************
The big winners were the taxi companies able to exploit an artificially high price of medallions to extract huge concessions from the local supply of drivers.
Uber et al. have not so much disrupted this value chain as they have created a parallel value chain where the end product wins on customer experience.
- Faster pick-up times, often less than 5 min.
- Superior coverage density that includes areas where dedicated taxi services are simply not viable.
- Payment simplicity.
- Civility encouraged by the dual-rating system. (Drivers rate passengers, passengers rate drivers.)
- Safety because of the auditability of rides.
Better. Again the suppliers are arguably getting exploited, but likely no more than they would be as traditional taxi drivers. (A labor market most Uber drivers are legally barred from in any case.)
The Uber et al. channel required a proliferation of mobile devices; but since they're here now,
transactions are cheap AND omnipresent, and the battleground again shifts to the user experience. The #1 differentiator is coverage density. But there are also analogies to the Amazon.com differentiators: machine intelligence to optimize routing (I reject the premise that there's no room for significant improvement here), and the ratings database adds more value the bigger it gets.
Traditional car-for-hire mobility is dead. Long live car-for-hire mobility.
Mobile-era car mobility strategy
Stepping back, we can think of the car-for-hire market as one piece of a much larger car mobility market.
***************************
Car Manufacturer - Uber/Taxi - Customer
-----------------------------------------------
Car Manufacturer - Rental net - Customer
==========================
Car Manufacturer - Dealership - Customer
Car Manufacturer - Dealership - Customer
Car Manufacturer - Dealership - Customer
***************************
The customers are all buying car mobility. Rental networks and taxis have always competed to an extent; while dealerships have been so differentiated there was little use for such an abstract view. But the mobile-era punchline is now obvious.
***************************
Car Manufacturer - Uber/Taxi - Customer
-----------------------------------------------
Car Manufacturer - Rental net - Customer \
========================== _\
Car Manufacturer - Dealership - Customer -- Uber
Car Manufacturer - Dealership - Customer _/
Car Manufacturer - Dealership - Customer /
***************************
Uber
can do this because transactions are now cheap and omnipresent; and it
will do this anywhere and everywhere that it can go head-to-head with car ownership and win on the user experience. (Which begins and ends with TCO.)
With each new customer relationship, Uber gets more leverage; and eventually they'll have enough to follow Amazon's path
back into the supply chain, looking for more leaks to plug in the user experience. They'll find many of these leaks in their physical fleet. And just like Amazon, Uber will move from the world of bits into the world of atoms---with considerable momentum.
At first, maybe its just a matter of tweaking the fleet composition. Perhaps user preference---for higher safety ratings, or lower emissions, or more luxury, or whatever---is being underserved in certain locales. Faced with the weakening demand at dealerships (amplified by the damage that lower EV service revenues inflict on the dealers' fragile cost structure), car manufacturers will at some point begin to negotiate large fleet deals directly with Uber.
Then Uber will really start to flex, pushing for its manufacturers to build cars that are
optimized for fleet membership. This will include a range of hardware redesigns. But much more importantly, it will mean software optimization, via machine learning, that can ONLY be done by direct integration with Uber's vast data assets.
The full consequences of this are tied into my next assumed trend, which I'll probably write up tomorrow.
* H/t to
Ben Thompson for a lot of these general ideas.
Last edited by Subfallen; 10-17-2015 at 02:07 PM.