In his, to my mind, fair defense of Uber, Mark Suster made a very important observation about the reality of business:
Let’s put this into perspective. As somebody who has to rub shoulders with big tech companies often I can tell you that there is much blood spilled in the competitive trenches of Apple, Twitter, Facebook, Google and so on. Changes to algorithms. Clamping down on app ecosystems. Changing how third-parties monetize. Kicking ecosystem partners in the nuts.
It’s a brutally competitive world out there because there are extreme amounts of money at stake. I’ve been on the sharp end of it and it doesn’t feel nice. And I pick myself back up, dust off and think to myself that I need to think through the realpolitik of power and money and competition and no matter how unpleasant it is – it’s a Hobbesian world out there. It ain’t pretty – but it’s all around us.
This is particularly relevant to Uber: the company is looking to raise another $1 billion at a valuation of over $30 billion, and, as I wrote when the company raised its last billion, they are likely worth far more than that. Still, though, skeptics about both the size of the potential market and the prospects of Uber in particular are widespread, so consider this post my stake in the ground1 for why Uber – and their market – is worthy of so many sharp elbows. I expect to link to it often!
There are three perspectives with which to examine the competitive dynamics of ride-sharing:
- Ride-sharing in a single city
- Ride-sharing in multiple cities
- Tipping points
I will build up the model that I believe governs this market in this order; ultimately, though, they all interact extensively. In addition, for these models I am going to act as if there are only two players: Uber and Lyft. However, the same principles apply no matter how many competitors are in a given market.
Ride Sharing in a Single City
Consider a single market: Riderville. Uber and Lyft are competing for two markets: drivers and riders.
There are a few immediate takeaways here:
- The number of riders is far greater than the number of drivers (far greater, in fact, than the percentage difference depicted by this not-to-scale sketch)
- On the flip side, drivers engage with Uber and Lyft far more frequently than do riders
- Ride-sharing is a two-sided market, which means there are two places for Uber and Lyft to compete – and two potential opportunities for winner-take-all dynamics to emerge
It’s important to note that drivers in-and-of-themselves do not have network dynamics, nor do riders: Metcalfe’s Law, which states that the value of a network is proportional to the square of the number of connected users, does not apply. In other words, Uber having more drivers does not increase the value of Uber to other drivers, nor does Lyft having more riders increase the value of Lyft to other riders, at least not directly.
However, the driver and rider markets do interact, and it’s that interaction that creates a winner-take-all dynamic. Consider the case in which one of the two services – let’s say Uber – gains a majority share of riders (we’ll talk about how that might occur in the next section):
- Uber has a majority of riders (i.e. more demand)
- Drivers will increasingly serve Uber customers (i.e. more supply)
- More drivers means that Uber’s service level (i.e. car liquidity) will improve
- Higher liquidity means that Uber has a better service, which will gain them more riders
In this scenario Lyft by necessity moves in the opposite direction:
- Lyft has fewer riders (i.e. less demand)
- Drivers will face increasing costs to serve Lyft riders:
- If there are fewer Lyft riders, than the average distance to pick up a Lyft rider will be greater than the average distance to pick up an Uber driver; drivers may be better off ignoring Lyft pickups and waiting for closer Uber pickups
- Every time a driver picks up a rider on one service, they need to sign out on the other; if the vast majority of rides are with one service, this, combined with the previous point, may make the costs associated with working for multiple services too high2
- Drivers will increasingly be occupied serving Uber customers and be unavailable to serve Lyft customers (i.e. less supply)
- Fewer drivers means that Lyft’s service level (i.e. car liquidity) will decrease
- Lower liquidity means that Lyft has an inferior service, which will cause them to lose more riders
The end result of this cycle, repeated over months, looks something like this:
There are three additional points to make:
- It doesn’t matter that drivers may work for both Uber and Lyft. If the majority of the ride requests are coming from Uber, they are going to be taking a significantly greater percentage of driver time, and every minute a driver spends on a rider job is a minute that driver is unavailable to the other service. Moreover, this monopolization of driver time accelerates as one platform becomes ever more popular with riders. Unless there is a massive supply of drivers, it is very difficult for the 2nd-place car service to ever get its liquidity to the same level as the market leader (much less the 3rd or 4th entrants in a market)
The unshaded portion of the “Riders” pool are people who regularly use both Uber and Lyft. The key takeaway is that that number is small: most people will only use one or the other, because ride-sharing services are relatively undifferentiated. This may seem counterintuitive, but in fact in markets where:
- Purchases are habitual
- Prices are similar
- Products are not highly differentiated
…Customers tend to build allegiance to a brand and persist with that brand unless they are given a good reason to change; it’s simply not worth the time and effort to constantly compare services at the moment of purchase3 (in fact, the entire consumer packaged goods industry is based on this principle).
In the case of Uber and Lyft, ride-sharing is (theoretically) habitual, both companies will ensure the prices are similar, and the primary means of differentiation is car liquidity, which works in the favor of the larger service. Over time it is reasonable to assume that the majority player will become dominant
I briefly mentioned price: clearly this is the easiest way to differentiate a service, particularly for a new entrant with relatively low liquidity (or the 2nd place player, for that matter). However, the larger service is heavily incentivized to at least price match. Moreover, given that the larger service is operating at greater scale, it almost certainly has more latitude to lower prices and keep them low for a longer period of time than the new entrant. Or, as is the case with ride-sharing, a company like Uber has as much investor cash as they need to compete at unsustainably low prices
In summary, these are the key takeaways when it comes to competition for a single-city:
- There is a strong “rich get richer” dynamic as drivers follow riders which increases liquidity which attracts riders. This is the network effect that matters, and is in many ways similar to app ecosystem dynamics (developers follow users which which increases the number and quality of apps which attracts users)
It doesn’t matter if drivers work for both services, because what matters is availability, and availability will be increasingly monopolized by the dominant service
Riders do not have the time and patience to regularly compare services; most will choose one and stick with it unless the alternative is clearly superior. And, because of the prior two points, it is almost certainly the larger player that will offer superior service
Ride Sharing in a Multiple Cities
It is absolutely true that all of the market dynamics I described in the previous section don’t have a direct impact on geographically disperse cities, which is another common objection to Uber’s potential. What good is a network effect between drivers and riders if it doesn’t travel?
There is, however, a relationship between geographically disperse cities, and it occurs in the rider market, which, as I noted in the previous section, is the market where the divergence between the dominant and secondary services takes root. Specifically:
- Pre-existing services launch with an already established brand and significant mindshare among potential riders. Uber is an excellent example here: the company is constantly in the news, and their launch in a new city makes news, creating a pool of riders whose preference from the get-go is for Uber
Travelers, particularly frequent business travelers, are very high volume users of ride-sharing services. These travelers don’t leave their preferences at home – when they arrive at an airport they will almost always first try their preferred service, just as if they were at home, increasing demand for that service, which will increase supply, etc. In this way preference acts as a type of contagion that travels between cities with travelers as the host organism
Most important of all, though, is the first-mover effect. In any commodity-type market where it is difficult to change consumer preference there is a big advantage to being first. This means that when your competitor arrives, they are already in a minority position and working against all of the “rich get richer” effects I detailed above.
This explains Uber and Lyft’s crazy amounts of fundraising and aggressive roll-out schedules, even though such a strategy is incredibly expensive and results in a huge number of markets that are years away from profitability (Uber, for example, is in well over 100 cities but makes almost all its money from its top five). Starting out second is the surest route to finishing second, and, given the dynamics I’ve described above, that’s as good as finishing last.
What I’ve described up to this point explain what has happened between Uber and Lyft to-date. Still, while I’ve addressed many common objections to Uber’s valuation in particular, there remains the question of just how much this market is worth in aggregate. After all, as Aswath Damodaran, the NYU Stern professor of finance and valuations expert detailed, the taxi market is worth at most $100 billion which calls into question Uber’s rumored $30 billion valuation.
However, as Uber investor Bill Gurley and others have noted, Damodaran’s fundamental mistake in determining Uber’s valuation is to look at the world as it is, not as it might be.4 Moreover, this world that could be is intimately tied to the dynamics described above. I like to think of what might happen next as a series of potential tipping points (for this part of the discussion I am going to talk about Uber exclusively, as I believe they are – by far – the most likely company to reach these tipping points):
- Tipping Point #1: Liquidity is consistently less than 5 minutes and surge pricing is rare – Once Uber becomes something you can count on both from a time and money perspective, rider behavior could begin to change in fundamental ways. Now, Uber is not just for a business meeting or a night out; instead, Uber becomes the default choice for all transportation. This would result in dramatically increased rider demand, resulting in complete Uber domination of driver availability. This would have several knock-on effects:
- Driver utilization would increase significantly, increasing driver wages to a much more sustainable level
- Competitor liquidity would decrease precipitously, leading to rider desertion and an Uber monopoly; this would allow Uber to raise rates to a level that is more sustainable for drivers, further increasing supply and liquidity
By all accounts Uber is already close to this level in San Francisco, and there are lots of anecdotes of people all but giving up cars.5 The effect of this change in rider behavior cannot be overstated, especially when it comes to Uber’s potential valuation: taxis have a tiny share of the world’s transportation market, which means to base the company’s valuation on the taxis is to miss the vast majority of Uber’s future market opportunity
Tipping Point #2: Uber transports not just people – Uber has already done all kinds of experiments with delivering things other than people, including Christmas trees, lunch, a courier service, even drugstore items. However, any real delivery service would need to have some sort of service-level agreement when it comes to things like speed and price. Both of those rely on driver liquidity, which is why an Uber logistics service is ultimately waiting for the taxi business to tip as described above.
However, once such a delivery service is launched, its effect would be far-reaching. First, driver utilization would increase even further, particularly when it comes to serving non centrally located areas. This would further accentuate Uber’s advantage vis-à-vis potential competitors: Uber service would be nearly instant, and drivers – again, even if they nominally work for multiple services – would be constantly utilized.
Moreover, there is a very good chance that Uber could come to dominate same-day e-commerce and errands like grocery shopping: most entrants in this space have had a top-down approach where they set up a retail operation and then figure out how to get it delivered; the problem, though, is that delivery is the bottleneck. Uber, meanwhile, is busy building up the most flexible and far-reaching delivery-system, making it far easier to move up the stack if they so choose. More likely, Uber will become the delivery network of choice for an ecosystem of same-day delivery retailers. Needless to say, that will be a lucrative position to be in, and it will only do good things for Uber’s liquidity.
Why Uber Fights
The implications of this analysis cannot be underestimated: there is an absolutely massive worldwide market many times the size of the taxi market that has winner-take-all characteristics. Moreover, that winner is very unlikely to be challenged by a new entrant which will have far worse liquidity and an inferior cash position: Uber (presuming they are the winner) will simply lower prices and bleed the new entrant dry until they go out of business.
To put it another way, I think that today’s environment where multiple services, especially Lyft, are competing head-on with Uber is a transitional one. Currently that competition is resulting in low prices and suppressed driver wages, but I expect Uber to have significant pricing power in the long run and to be more generous with drivers than they are now, not for altruistic reasons, but for the sake of increasing liquidity and consistent pricing.
In short, Uber is fighting all out for an absolutely massive prize, and, as Suster noted, such fights are much more akin to Realpolitik. As Wikipedia defines it:
Realpolitik is politics or diplomacy based primarily on power and on practical and material factors and considerations, rather than explicit ideological notions or moral or ethical premises
It’s ethics – or, to be more precise, Uber’s alleged lack of them – that has been dominating the news most recently, and is what inspired Suster’s post. And, to be very clear, I can understand and share much of the outrage: in my Daily Update I have compared Uber to Wall Street and said that Emil Michael should be fired (both links members-only) for his comments suggesting Uber might investigate journalists – Sarah Lacy in particular – who disparage the company.6
However – and one of the reasons I’m writing this article – I am also very aware of just how much is at stake in this battle. Lyft has raised $332.5 million from some very influential investors, and I don’t for a minute believe that they don’t want to win just as badly as Uber does. It’s perfectly plausible, if not probable, that Lyft and its backers, overmatched in a head-on battle with Uber, are conducting a guerrilla campaign with the aim of inspiring so much disgust in riders that Uber’s liquidity advantages start to slip (and to be clear, such a campaign – if it exists – is only possible because Uber’s management speaks and acts poorly frequently).7
To be perfectly clear, I don’t know anything further about this situation – or other recent Uber PR fiascos, like this Verge piece about stealing Lyft drivers – beyond the size of the potential prize, as detailed here, and the reality of human beings and their incentives in the presence of such outsized rewards. In my experience the truth ends up being far more gray than the press – which really hates threats to journalists – has characterized this most recent episode.
In fact, in some ways I’m actually far more concerned about Uber’s perceived lack of ethics than most, because if I’m right, then Uber is well on its way to having monopoly power over not just taxi services but a core piece of worldwide infrastructure, and nothing about this crisis gives me confidence in the company’s ability to manage that gracefully.8 I get that Uber’s willingness to fight unjust laws is what got them to this point, but as James Allworth and I discussed on the most recent episode of Exponent, there is a deeper moral code that ought to govern Uber’s actions. Moreover, Uber needs rider goodwill to prevail in the many markets where it is facing significant regulatory resistance: it is local citizens who determine whether or not local laws and regulations will be changed to accommodate Uber, and Uber is making it very difficult to rationalize advocating for them, or, if my Twitter account is any indication, to even ride with them.
Ultimately, this blog generally seeks to analyze business, not render moral judgment or tell anyone what products or services they should or should not use. I myself am mixed: I plan on spending some time in the white part of that graph above, at a minimum. I hope, though, that you now appreciate exactly what is at stake and why so many elbows are being thrown.
Originally, this bullet stated “Drivers will increasingly be occupied serving Uber customers and be unavailable to serve Lyft customers (i.e. less supply).” However, this was incorrect because drivers utilized on any service are unavailable to every service, incurring no advantage ↩
Lots of people have suggested to me that Uber will be doomed as soon as someone creates an app that serves as a front-end to all of the services allowing you to book the one with the lowest price and/or fastest availability; however, such an app would realistically need the cooperation of the largest player (which would not be forthcoming, and there is no public API) plus need to gain meaningful traction in a given market while competition still exists. It’s not happening ↩
The broader effects of Uber on adjacent industries will have to wait for another post ↩
That said, no reporting has suggested a threat to Lacy or her family as many seem to believe; that came from Lacy herself ↩
First off, Michael’s comments, whether in jest or not, were incredibly stupid. Secondly, Kalanick’s tweetstorm was a terrible idea. You can’t admit that Michael’s “remarks showed a lack of leadership, a lack of humanity, and a departure from our values and ideals” and not fire him. Either stand your ground and insist Michael was misrepresented or let him go ↩