Tuesday, December 2, 2014

Up, up and away! A crowd-valuation of Uber!

In June 2014, I tried to value Uber and arrived at an estimated value for the firm of $6 billion, an impressive number for a young firm, but well below the VC estimates of value of $17-$18 billion at the time of my post. Much of the reaction was predictable, with readers whose priors were confirmed by my assessment of value liking it and those whose priors were different disagreeing,and sometimes vehemently. Disagreement and debate don't bother me in the least, since they can only advance the valuation narrative, but I do think that putting my narrative and valuation front and center undercut my objective in two ways. First, it made for passive analysis, where you could pick and choose which one of my valuation inputs you agreed with and which ones that you found erroneous, the justifying your prior biases. Second, some who disagreed took the easy way out, arguing that it was my use of an intrinsic value (DCF) model that had led me down the wrong path and that it was therefore unfixable. 

Now that Uber is in the news again, with value estimates of $40 billion and higher floating around, I decided to revisit the valuation, but from a different angle. Rather than presenting my valuation, I want to open the process up and I would like to invite you along for the journey. Like a book or movie where you get to write not just the ending but the entire story, I will provide the architecture and you can build your own valuation story (and value) for Uber. The good news is that this valuation will reflect your views (not mine) on Uber. The bad news is that if you don't like the value, you cannot blame me.

When Narrative drives Value
While my original valuation of Uber was all about the numbers, I followed it up with a post where I argued that if you disagreed with my value, it was not because you had a problem with my estimates (of growth or risk) but because you were taking issue with my narrative. Underlying my original valuation was a story that I was telling about Uber as an urban cab/limo service company that would continue to attract new users into the market, while maintaining its high profit margins. In response to a post by Bill Gurley, venture capitalist investor (and director) in Uber, where I was accused of missing the story by a mile, I conceded that I knew far less than he did about the company and that his narrative for the company - Uber as a car-service for the masses with global networking benefits - would lead to a much higher value for the company

While that may sound abstract, the best way to see the link between story telling and number crunching is to take Uber on the valuation process, with you making your judgments at each step of the way. As you make this journey, a few (gentle) reminders of issues that you will face along the way:
  1. This is your valuation: Contrary to what you might have been taught in your valuation classes, valuations are and should never be just about the numbers. To the extent that you will be making choices on these number, this will be your estimate of valuation, reflecting not only what you know about the company (and its products, management etc.) but also your personal biases (whether you like the company or not). 
  2. You are almost certainly wrong: Lest you view this is an insult, so is my assessment of value and so are the VC’s valuations. It is not because we don't understand valuation or have not done our homework, it is because we are trying to play God and forecast the future. 
  3. You should be open to revisiting it: Following up on the last proposition, it stands to reason that the choices you make in valuing Uber today will not be the choices that you will make tomorrow or a week from now. So, keep the door open for changes not just at the margins but in your central narrative.
  4. Be willing to act on it: There is no point to valuing companies, if you are not willing to act on your valuations. With Uber, it is true that you and I are restricted in what we can do, since the company is still private. However, it is also clear that the explosive growth in the estimated value of the company sets it on a path to being public (sooner, rather than later), at which point our valuations will become actionable.
Setting the stage
The first step in valuation is assessing where the company is right now and we start off at a disadvantage, because it is amazing how little we know about the operating details of a company that is in the news as much as Uber. According to the company's website, it operates in 51 countries and in about 230 cities on six continents, and it has also expanded its product offerings, both within the car service market (with U4B, directed at businesses and UberPool, allowing for car pooling)and in new markets (with UberRUSH, its delivery service in New York City).

The only updated revenue numbers came from an article in Business Insider, which seems to be one the company's preferred venues for leaking selective information. According to the article, the company projects gross receipts of $10 billion in 2015, up three times from gross receipts in 2014, which in turn more than tripled relative to receipts in 2013. While the company originally kept 20% of these receipts as revenues, it is unclear whether that number has slipped in recent months, as it has gone aggressively for new growth. While I am normally loath to value companies based upon second-hand information, and especially so if the information comes from a leaked corporate document, I am going to assume that the company will generate $3.5 billion in gross receipts for 2014 and that its slice has stayed at 20%, giving it revenues of $700 million for the year. I have no idea whether it is profitable after covering its operating costs, but the impact on the final value of these initial numbers is small enough that it is worth moving forward.

Building your Uber narrative
To set up the link between the narrative that you will be telling for Uber and its value, I will borrow the set-up that I used in this post on narrative and numbers, where I took the key inputs into my valuation and connected them to stories told about companies:


There are thus six steps to the narrative process and your choices at each step will determine the numbers from which we estimate value.


Step 1: Potential Market
In my initial valuation of Uber, I treated it as an urban car-service company and was taken to task rightly for having too cramped a vision of the company. It is quite clear from both its words and actions that Uber has much larger designs and I will leave it to your judgment whether it will succeed. Based on rudimentary research of the potential markets that Uber could be in, here is what I get as a list:

The potential starting market can range from $100 billion (for urban car service) to close to $300 billion (if you treat it as transportation company, going after all of the markets above). Since this is your narrative, its your choice to make and it will have significant value consequences. 
Based on what you know (and think about) Uber, which of the following do you think is its potential market?

Potential Market
Market size (in millions)
Description
A1. Urban car service
$100,000
Taxi cabs, limos & car services (urban)
A2. All car service
$150,000
 + Rental Cars+ Non-urban car service
A3. Logistics
$205,000
 + Moving + Local Delivery
A4. Mobility Services
$285,000
 + Mass Transit + Car Sharing


Step 2: Market Growth
Uber is not only disrupting the existing players in the market that it disrupts but it is also attracting new users into the market, either by attractive non-cab users to try Uber or increasing the usage of car services, in general. Assuming that this process continues, the growth rates in these markets could increase if Uber's services (or Uber-like services) become more widely accessible. Here again, the choice is yours.
Based on the potential market(s) you chose for Uber in step 1, what effect do you see Uber (and Uber-like services) having on the expected growth rate in the market?

New user effect on market growth
Annual growth rate (next 10 years)
B1. No new users (no growth effect)
3.00%
B2. Increase total market by 25% over next 10 years
5.32%
B3. Increase total market by 50% over next 10 years
7.26%
B4. Double market size over next 10 years
10.39%

Step 3: Market Share
Having chosen a potential market and a growth rate in that market, the third step is making a judgment on what market share you would expect Uber to command once the market hits steady stay (in ten years). That choice will depend in large part on whether you think Uber's products/services have network effects, where increased usage of Uber by customers in a market makes it more attractive to other potential customers, and whether you think these network effects are local (in the city/region of usage) or global (in other cities/regions). The arguments for local network effects are easy (the more Uber users there are, the more Uber cars there are, which in turn makes it easier/quicker to get an Uber ride)  but the ones for global network benefits may be more of a stretch (links to credit cards, inertia, uniformity of service, staying with the known). Once you have assessed the pluses and minuses, here are your choices.
Based on your assessment of Uber, what type of network effect (if any) do you see for its products and services?

Network Effect
Market Share
Description
C1. No network effects
5%
Open competition in every market
C2. Weak local network effects
10%
Dominance in a few local markets
C3. Strong local network effects
15%
Dominance in multiple local markets
C4. Weak global network effects
25%
Weak spillover benefits in new markets
C5. Strong global network effects
40%
Strong spillover benefits in new markets

Step 4: Revenue Slice & Operating Costs
Uber gets to keep a portion of the gross receipts paid by users for an Uber service, representing their revenues. That slice was initially set at 20% of the receipts but whether it can stay at that level will depend upon both the markets that Uber decides to operate in and the competition within each market. Thus, if Uber decides to go into the logistics market (moving and local delivery), it will have to accept a much lower slice of revenues, since competition is more intense. Even within the urban car service market, more intense competition from existing players (Lyft) or new entrants could put Uber's revenue slice under pressure. This choice again is yours to make:
Given the markets that you see Uber entering and the competition it faces within those markets, how strong and sustainable are Uber's competitive advantages?

Competitive Advantage
Revenue Slice
Description
D1. None
5%
Unrestricted entry + No pricing power
D2. Weak
10%
Unrestricted entry+ Some Pricing Power
D3. Semi-strong
15%
Unrestricted entry + Pricing Power
D4. Strong & Sustainable
20%
Restricted entry + Pricing Power

Step 5: Reinvestment Needs
Uber's existing business model, where it acts as an intermediary and does not invest in cars or equipment, has low capital intensity and as a consequence, much of its growth has come with relatively low reinvestment. That could change, if Uber decides to change its business model or if it has to do acquisitions to continue to generate growth. 
Based on the business model that you see Uber adopting as it goes for the market share (that you forecast) in your potential market, which of the following reinvestment policies best fits the company?

Reinvestment
Sales/Capital Ratio (Higher number= Less investment)
E1. Minimal capital needs, no acquisitions
10.00
E2. Minimal capital needs, small acquisitions
5.00
E3. Service company median
3.00
E4. Technology company median
2.50
E5. US company median
2.00
E6. Capital intensive company median
1.50

Step 6: Risk (Cost of capital & Survival risk)
As I noted in the table above, there are types of risk that you have to grapple with in valuation. The first is the risk in operations, which causes revenues and earnings to be volatile over time, and that risk is captured in the risk-adjusted return you demand for investing in the company. In valuation, the cost of capital becomes the measure of this risk-adjusted return and is generally estimated by looking at publicly traded companies (even though Uber is privately held still). Rather than wrestle with the minutiae of inputs into the model, you can make a judgment on where in the cross-sectional distribution of costs of capital across all companies you would put Uber.
Based on your assessment of the risk in the market that Uber is entering and where the company is in its life cycle, what cost of capital would you pick for the company?

Risk Profile
Cost of Capital
F1. Lowest decile of US companies
7.00%
F2. 25th percentile of US companies
7.50%
F3. Median of US companies
8.00%
F4. 75th percentile of US companies
10.00%
F5. Ninth decile of US companies
12.00%
The other risk for a young company is survival risk, i.e., the risk that you are one disaster from shutting operations. That risk increases for smaller companies with small cash holdings, large cash needs and limited access to capital. Given Uber's capacity to raise capital and cash holdings, this risk should be lower.

Your Uber value
Once you have made the choices on the potential market, growth in that market, Uber's market share and revenue slice, the valuation follows. While the number of combinations of assumptions is prohibitively high to show value estimates under each one, I have summarizes the value estimates for at least a subset of plausible choices. (using a sales to capital ratio of 5.00 and a cost of capital of 12% for all the cases)>

If your set of assumptions is not listed above, you can download the spreadsheet, enter your choices and see what the value of Uber is with those choices. If you don't like the value that you get with your narrative choices, I am afraid that it is just a reflection of your choices.

Looking at the range of values that you can obtain ($799 million to $90.5 billion), you may find your worst fears about DCF models, i.e., that they can be used to deliver whatever number you want, vindicated, but that is not the way I see it. Instead, here are four lessons that I draw from this table:
  1. Soaring narratives, soaring values: I know that some people view DCF models as inherently conservative and thus unsuited to valuing young companies with lots of potential. As you can see in the table above, if you have a soaring narrative of a huge market, a dominant market share and hefty profit margins, the model will deliver a value to match. Put differently, if you found my original valuation of Uber too low, the fault lies with me for having a cramped vision of what Uber can accomplish and not with the model. It also stands to reason that when you have big differences in value estimates, it is almost always because you have different narratives for a company, not because you have a disagreement on an input number.
  2. Not all narratives are made equal: While I have listed out multiple narratives, some of which deliver huge values and some not, not all are equal. Looking forward as investors, some narratives are more plausible than others and thus have better odds of succeeding. Looking back ten years from now, reality will have delivered its own story line for Uber and the narrative that came closest to that reality will be the winner.
  3. Narratives need reshaping: The narratives for Uber that you developed are based on what you know today. As events unfold, it is critical that you check your narrative against the facts and tweak, change or even replace the narrative if the facts require those adjustments, which was the point that I made in this post.
  4. Narratives matter: Success, when investing in young companies, comes from getting the narrative right, not the numbers. That may explain why some successful venture capitalists can get away being surprisingly sloppy with they numbers. After all, if your skill set includes finding start-ups with strong narratives and picking founders/entrepreneurs who can deliver on those narratives, the fact that you cannot tell the difference between EBITDA and free cash flow or compute the cost of capital will be of little consequence. 
If you are waiting for me to reveal my narrative choices, you will be disappointed. This is your valuation, not mine, and I hope that you like it. If you could please go in and put your narrative choices and resulting value for Uber into this shared Google spreadsheet, we can get a crowd valuation of Uber!

Attachments

17 comments:

jk0621 said...

Hi professor,

Always a fan of your work and really impressed you took on the challenge/controversy of valuing things like FB, TWTR, TSLA and AAPL in the past and were pretty good at trading them too!

On Uber, I would think that a bottom-up way to estimate revenue would be better than looking at the overall market - since it's hard to say whether its going to enlarge the market by 50% or 300% (and I believe this is probably the most relevant narrative debate - whether this is habit changing or just market share gaining).

We can assume a certain $/trip (and even break it out by city). For convenience I will assume a $15 gross receipt, with higher numbers in NY, SF, London but assuming smaller tickets in EM. A 20pct cut off of that would be ~3$/ride.

For quantity, we should focus I think on high density urban dwellers since this would likely be the main market. I have excluded very underdeveloped EMs, as well as China. I believe in China Uber will be limited to foreign travellers and local players like Tencent's Wechat Taxi service is likely to dominate (this has been the story again and again for China, a foreign company gets max 10% share). I will just exclude the China opp to be conservative. This gives me a market size of about 600m people. Of this TAM, I will assume 1/3 of them use the Uber service.

Rides / person / year is the biggie. Someone who travels every day via Uber will average 700 rides / day while some may only use the service once a month or a few times a year. I think 5% of the users or 9.9m people globally can easily use the service 200 times/year on average. Another 10% or ~20m may use on a weekly basis and average 50 trips / year. Another 20% I will assume uses 10 trips per year and the rest I will assume only use 5 trips per year on Uber.

This will lead to total revenues of about ~$12bn, on assumption of $3 net revenue per ride (gross billing of $15). If Uber can manage a 30% profit margin (light infrastructure costs, network effects) this would lead to $3.6bn revenue at maturity (maybe 10 years from now) - and giving a stable, platform like multiple of 18x (Google, Visa come to mind) would give it a valuation of $67bn at relative maturity.

I think these assumptions are reasonable and somewhat conservative, esp on rides / user and $ / ride, although perhaps a bit aggressive on percentage of population that use the service.

Tom said...

Hello Prof.,

one of the best posts of recent months. Not because it provides with shocking new insights, but because you take the time to explain what the different choices could be and why and why it is important to have a look at the narrative.

The Excel file is then a "piece of art", as it allows playing with the inputs, and while there are many different view on what might or might not be important, it is a consistent toy to play with (especially as it allows manual inputs at times)!

What would be interesting to lean how you go about getting your inputs on market size etc.

Please keep up the excellent work!

Anonymous said...

Is it possible you take this to the next level and articulate how one can arrive at a set of internally consistent numbers for the different narratives? What checks do you do? What sources of data do you depend on?

Frank said...

Great angle to take.
It's too easy to sit on the sidelines and debate inputs.
Respect goes to anyone who has the courage to actually play the game, regardless of relative success.

Aswath Damodaran said...

Tom,
On the market size, I take what I can get and it often means going to many sources. Here are some that I find useful:
1. Industry trade groups: Almost every business has a trade group that reports aggregate statistics. For instance, to get the data on the size of the moving services business, I went to to the American Moving & Storage Association (AMSA) website and they do an annual update of the details of their business.
2. Business forecasting services: There are services that estimate the business size and provide forecasts for the future. Here is one of my favorites (IBISWorld) and it provides details on a vast array of businesses. Two problems. First, he data is often better for the US than global. Second, most of these services require payments (and some are pretty significant). So, I take what I can get at low or no cost. I also try to find multiple sources of data so that I can crosscheck the data. (That requires a couple of hours of focused Google search).
3. Company level data: I have access to Capital IQ which has data on both public and private companies (>100,000 globally) broken down by industry to the nth detail. I download the revenues of every company in a business and cross check against my aggregate values.
In spite of all of these efforts, you may still find your self faced with uncertainty, as I was with the cab business (where I had to go city by city and then extrapolate from there).

Aswath Damodaran said...

Anonymous,
You are right. Narratives have to be checked
(a) Against history: Has any company done what you are forecasting for your company before? If yes, how does your company measure up against that company? If no, what will make it the first?
(b) Against Economic first principles: If you are providing a high margin/high revenue growth narrative, how do you reconcile with the fact that higher prices usually translate into lost revenues?
(c) For consistency: Are your assumptions about risk, growth and cash flows internally consistent? In other words, are you assuming high growth with low/no reinvestment and low risk (which would be inconsistent)?
I will work on a post specifically on this issue or weave into a different valuation.

Anonymous said...

Off topic, any ideas on the current state of US independent oil and Gas companies. They been crushed lately, possible opportunities available??

Vikrant Mahajan said...

Hi Prof,

Though i am a software professional, i have taken to Valuation and corporate finance as a hobby class after my 1 year exec education programme from IIM Bangalore in early 2013. I have been following your blogs since then and they are not just thoughts but classes in themselves. Thanks for such insightful articles.

Having said this, i wanted to say that i not only acknowledge the market size that you have considered but i feel that it is still an optimistic figure. One important factor that everyone seems to be missing is the fact that governments especially in emerging markets are looking to provide highly safe and best in class public transportation to decongest the populated roads. Good quality Public transportation is in itself a competition to taxi services like Uber who thinks that networking effect will make people travel more in taxis. Metro in New Delhi has been a big hit and lot of people from lower to upper middle class prefers it more than taxis, not only for cost factor, but security, comfort and punctuality in reaching the destination. Nobody wants to get stuck up in clogged roads of heavily populated areas in Delhi or bangalore and alike cities. Most people take taxis only if they are in real hurry OR if they have to goto Airports or railway stations and for that mom n pop taxi stores are still a good bet.

Moreover, the recent incident in Delhi and subsequent ban reinforces the fact that road for Uber in such countries where there are no security and identification f/ws in place is going to be ONLY hard and not easy. Uber cannot just claim to be a tech pltform BUT has to take ownership of driver profile verification PLUS a robust GPS system that does not just reply on PHONE and an installed APP.

I have shifted base to US in VA in the month of April 2014 and hence could not test the service in India but based upon my discussions and thoughts in papers, I think the road ahead for Uber is just too difficult in a market like India not only due to security reasons BUT also a great competitor in the form of METRO which is expanding at a brisk pace in entire Delhi and nearby cities.

What is your viewpoint on these thoughts ?

Anonymous said...

I just caught a snippet of your comments re Uber tonight (Dec 10, 2014) on Nightly Business Report. Did you change your view re Uber as you said on the program that Uber still had a lot of room to grow? Or do you still think Uber's valuation is absolutely crazy? Thanks.

Aswath Damodaran said...

You can believe that Uber has lots of room to grow and believe that it is over valued at $40 billion, at the same time.

Yniim said...

Hi professor,

Given that there is a big risk that Uber may be banned in some countries, how do you factor that risk in your valuation analysis ?

Naina said...

Is the $100 Bn size of the taxi and limo market only for US cities, or for the global market?
Any data on % revenue outside the US for Uber?

TimurIV said...

I think in the future it would be very convenient to have a fleet of driver-less cars. I am not sure how Uber would factor into this...

Colby's Dad said...

Feel free to ignore, but suggest you change the "Goal Seek" picture in the "ERPJan15" excel file as it took me a little bit to understand the source of the goal seek amount. Pretty sure it is the picture from January 2014. Otherwise, thank you for another great post.

FIPÃO said...

Hi Professor,
I think the food delivery market is more reasonable to measure, therefore Just Eat (ticker: JE/:LN) is being negotiated ~3 times TPV, and Uber around 4 x TPV (in a much larger industry and with more cross selling opportunities).What are your thoughts about just eat valuation? Many thanks!

Matias Orellana said...

Dear Professor,
I sympathize with your previous view on UBER's value and would like to thank you for sharing such interesting information. I wanted to point out some other difficulties that make it harder for UBER to expand globally.

On your previous UBER post, you mentioned that Latin America was a market that could grow beyond averages because it is under served. Actually this is not the case, for example in Peru, you have a total of 240k taxis on the market while the optimal amount is around 100k so there is an over capacity of 140k (good source) (Here it is said that taxis are the insurance of the unemployed). The same scenario can be seen in neighboring countries, so rather than growing markets, they could be diminishing ones, due to overcapacity, alternative transportation development and higher accessibility to car purchases.

But the biggest concern is competition. In South America the main taxi matchmaking company is EASYTAXI not UBER, they are quiet well positioned and it won’t be easy to take some of their share. I think the biggest reason that puts UBER on a disadvantage to other companies is that in order to ask for a cab you have to introduce your credit card data before. In some places, especially in underdeveloped countries, people prefer to pay in cash because of trust and technological issues. (This can be solved so it’s not very relevant).

Competition is also worrying on the new markets under consideration. In Europe for example, there are already car sharing apps such as BlaBlacar, which is the market leader by a great extent. In car rental biz, you have companies such as Drivy in France and the list goes on...

UBER is an amazing platform, the optionality to expand to other businesses is very wide, the market is big and growing, but despite all of it, the barriers to entry to these kinds of businesses are very low, this will eventually lead to a proliferation of competitors and margins will shrink.

I think you made it very clear on your prior post that the company was overvalued, even considering very favorable assumptions. Once again, enthusiasm won over rationality.

Eugene Weixel said...

I am a New York City fleet yellow taxi driver who's been following Uber for a while and blogging about it.

I don't know if you know that Uber inflates and distorts its driver income numbers in order to recruit drivers who usually are not sophisticated. If you scope out Facebook and other platforms you'll find groups of Uber drivers who are awakened to the scam. They work at dissabusing potential driver recruits and sporadically organize strikes and demonstrations against Uber.

Then you have legal challenges to the very core of their business model in the United States- the classification of Uber drivers as independent contractors. Super exploitation of desperate people is Uber's "secret sauce" and not any technological toys. There's a real possibility that this core asset that Uber has - super exploitation of desperate people trapped in subprime usurious car loans will come undone.

There is also the admittedly long shot possibility that e hail driver apps themselves will be banned as the public safety threat that they are. As it stands New York City has taken steps to finally minimize the problem of app and telephone distracted driving. This is an implicit admission that the problem exits. I know of two lawsuits against Uber over this situation in New York alone.

As a taxi driver I can report that it doesn't seem like Uber has hurt my income yet. There are fewer yellow taxis on the streets of the city but there's a whole armada of green taxis operating that are pretty new to the game.

I would take any figure emanating from Uber with a giant grain of salt.

By the way the recent leak that Uber has an operating deficit of $275 millions- do we even know if that number is reflecting losses for a year? A quarter? A month?