Posted on by meanreversionblog · 1 Comment

Ridesharing’s saturating usage problem!

When comparing the business drivers of both Uber and Lyft, two things stand out: Uber’s scale and its international reach (40% of revenues). Uber operates in 57 countries already.

But elsewhere they look very similar. Of-course to make any comparison we have to make assumptions around a) mix of pool rides being similar (as they are counted as separate trips/rides), and b) UberEats rides are not really that big as a share of total and don’t skew things much (its share of net revenue is ~5%; booking share 20%).

Drivers of revenue growth – Lyft and Uber

Ride-sharing businesses have relatively straight forward drivers here on: user growth and usage growth. Other inputs seem to be running at a level not leaving much room for growth. A take rate of 20-27% is already quite rich. And an average ride of ~$10 sounds very reasonable. Ofcourse, its worth nothing that this averages shared and single rides.

User growth is likely to continue for a long time, but its costly.

Software subscription business or businesses with repeat customer purchases, and in this case Uber and Lyft, spend cash upfront on sales & marketing to acquire new customers with the hope of keeping them over multiple years. In my estimate both Uber and Lyft spend ~20 months worth of gross profit on acquiring a new customer (in line with most such business assuming decent retention). This explains their very high Sales & marketing spending (28%, 37% of sales). Growing number of users is very costly!

Usage growth from existing customers is what funds new user growth.

Subscription businesses fund this costly growth of acquiring new customers by increasing the share of wallet of existing customers.

With respect to usage, I note two things from Uber and Lyft’s S-1 disclosures:

  • Uber users use it more often, ~50% more. There are two potential explanations for that – either Uber has a higher share of shared rides, or there is greater usage of ride-sharing outside of North America.
  • Usage per rider seems to be flat-lining for now. This is not good news, especially given there is potentially a TAM for 400 rides per rider (5 days a week * 40 weeks * 2x a day) rides every year. For some, this might mean that there is a huge TAM, as both Lyft and Uber would also suggest, and we are only 10-15% penetrated. Or something less benign …

The issue for ride-sharing is that usage and thus wallet share of existing users appears to be plateauing. Perhaps this is because the next best alternative for transportation, i.e. public transport, is still cheaper compared to ride-hailing. In London, a tube journey costs ~£2.8 one way, and a pool easily 2x more and takes longer. And a pool is not really more convenient versus public transport when you have to walk to designated pickup locations.

Flat-lining usage growth, adjusting for mix

What next? A tricky trade-off

Fully self driven cars will drastically reduce costs for ride-hailing services and likely lead to demand elasticity and open up the S-curve for further usage penetration gains. But robo-taxis, with no drivers, even by most optimistic estimates are at least 5-10 years away.

This leaves a tricky trade-off at the hands of these businesses:

  • Driving usage growth would require price declines and would mean expanding losses. With already cash-consuming user growth, this will require more capital – likely debt raise.
  • Accepting flat-lining usage growth means driving growth solely through growing users. This would lead to growth slowing down faster than what many expect.

Over the next 3-5 years, these businesses will need to choose between profitability and growth, and will likely require more capital. Something I think might happen earlier than what many expect.

This is not investment advice.

Posted on by meanreversionblog

Why does Facebook want regulation?

The backlash against the Tech giants has hardly been out of the headlines in the  past two years. Google has been fined twice by the European regulator, Facebook has appeared in front of the US Senate to explain its involvement with Cambridge Analytica and the European Parliament has passed a copyright directive, making internet platforms liable for content their users upload.

A tumultuous two years was followed last month by Mark Zuckerberg calling for regulation in an article for the Washington Post. But why does Facebook want regulation?

The debate

A year ago, I discussed that most of the issues with various stakeholders can be summarised into three categories: a) Anti-competitive practices b) Ownership of content, and c) Compromised user privacy.

While the issue of anti-competitive acquisitions is straight forward to fix with greater scrutiny, the last two issues are intertwined and at conflict with each other. 

These issues arise not only because these digital platforms are maximising their economic returns, but also because of laws written back in the mid-1990s to sustain growth of the Internet. In the US, the law suggested “No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.” Because that law applied only if the information service provider was not the publisher nor played any editorial role, the platforms responded to the incentives and remained neutral to user generated content. Similar laws were passed elsewhere. This allowed the advertising based business model for the internet to scale.

Given how the internet has developed since these laws were established, these rules are out of date and recent events have demonstrated that they can be abused by ‘bad’ actors – state sponsored, extremists or individuals.

So how do we set new rules? There are two issues that need to be addressed:

The issue of scale

Let’s tackle this one first.  Facebook and YouTube both reach roughly two billion users, and every second a huge amount of content is added to the existing vast content library they host. In the case of YouTube, 400 hours of content is uploaded every minute. The challenge for the tech giants is how to moderate such a huge amount of content and if it is humanly possible.

If the moderation is automated, what is the balance to ensure that the tech giants are not censoring free speech or debates on topics that are deemed offensive?

While automation will be key to the solution, alongside human judgement,  it is unlikely to be without error. This means that such a solution will not be available imminently for the long term as it would require several iterations to get the balance right.

Conflicting requests

Content regulation and user privacy are difficult to reconcile at the same time. While governments increasingly want the digital platforms to be responsible for the content they host, the users want more privacy and control over their data (i.e. content on these platforms). How do you satisfy them both?

Governments have wanted a degree of control over the content being shared on social platforms. To achieve this, their answer is to make social platforms responsible for the content they host. A failure to comply could lead to fines for the company and/or imprisonment of executives. At the same time, we have heard users and privacy groups voice their concerns about control that platforms have over user data. They want to liberate it. Regulators, on the other hand, want more competition and to break the monopoly of these platforms over user data. However, if and when the data is liberated, it could be exposed to misuse, as was the case with the Cambridge Analytica scandal, or it could perhaps increase the chances of it getting hacked and ending up in the hands of ‘bad’ actors. If the data is encrypted to ensure it ends up only in the right hands, control over the data is potentially lost, which is not what governments want.

So, where do we go from here?

Given the breadth and reach of these platforms today, the tech giants now have a responsibility to distinguish between good and bad user generated content. However, that could lead to potential errors of judgement and the censoring of free speech. 

For a wide reaching and mature internet, we need new rules. Governments need to think long and hard before writing those new rules, as they will decide the internet of the future. Regulatory burden in most industries entrenches the market position of incumbents and makes it onerous for new entrants to compete. The governments need to balance the ‘red tape’ with the level of competition in the industry and consumer surplus they want to preserve. Why does Zuckerberg want regulation? He knows the internet needs new rules and the trade-offs that will be involved to satisfy everyone. Anticipating the issues that will become apparent further down the road, he is asking for a unified set of requests from different stakeholders that get more complex when you involve multiple governments. Standardised regulation is probably in everyone’s interest – governments, consumers, new entrants, and Facebook.

Not Investment Advice. This blog first appeared at www.equitiesforum.com


Posted on by meanreversionblog

The wisdom of death

I was chatting with a friend about death this morning. She told me about a man on his death bed who told her what he would do right if he was given another chance at life – spend his time doing what he wanted!
A realisation hit me that nothing helps us prioritise (our lives) better than the realisation that we will die one day.
What better way to keep that realisation fresh everyday than this quote from Seneca (Letters from a Stoic):

Show me a man who reckons the worth of each day, who understands that he is dying daily? For we are mistaken when we look forward to death; the major portion of death has already passed.