Revenue Multiples And Growth
When you say “the stock is trading at 20x revenues” people rightly shake their head and say “that is nuts.” I got a lot of tweets like that in reaction to my comments about the Uber valuation in the LeWeb breakfast chat.
However, what people fail to realize is these things happen in a moment in time and that stocks won’t trade at 20x revenues forever.
Let’s take a fictional company that has $1bn in revenues in 2014 and goes public at $20bn, 20x revenues. Let’s say it will double revenues in 2015, then grow 60% in 2016, and 40% in 2017, and 30% in 2018.
So here are the revenue numbers
So, let’s now look at profits, since valuations are ultimately a function of profits, not revenues.
Let’s say this fictional company is breakeven in 2014, but expects to make 10% EBITDA margins in 2015, growing to 25% EBITDA margins by 2018. So here are the EBITDA numbers that fall out of that.
Let’s say this fictional company’s stock will go up 10% a year each year until 2018. So the valuation goes from $20bn today to $29bn over five years.
So here are the Revenue and EBITDA multiples that fall out of this thought exercise
The point of all of these numbers is to show that if a company can grow very quickly over a five year period, and become highly profitable, the stock can perform well and the multiples can come down to earth pretty quickly.
That is a bunch of “ifs”, but every once in a while this actually happens. It happened with Google which now trades at 5-6x revenues, and it is happening with Facebook which is in the middle of this kind of a story. So my point is 20x revenues is a huge number, but every once in a while, a company actually deserves it.
For anyone who wants to dig into the numbers a bit more, here’s a link to the google sheet that I built as I wrote this post.
Thanks for sharing. This is one of my favorite AVC posts in 2014.When I was a VC (for about a year after UBS) we mostly tried to look for investments that had 8-12X. Even those are exceptions. The 20X are really unicorns.Most companies operate around 3-5X times if that.The 20X decays (opposite of exponential growth) because once the majority of the initial market is captured, CAPEX is on scaling support for existing customers rather than new acquisitions and their contingent revenues.
The investment supply-revenue depletion curve is akin to Mundell-Fleming’s ISLM curve (for non-economists…it’s how the government does Investment Saving–Liquidity Preference Money Supply).Since Mundell-Fleming’s time we’ve discovered the curves are shaped less like Moore’s law exponential growth curve and more of an S-shape because there’s a gradation to a plateau where growth reaches steady state (fewer and fewer new markets and clients to capture).The consideration for investors is where on the S-curves (investment growth for new acquisitions and its counter-curve of depleting demand) to invest so that the gradient for ROI is steepest, aka before the plateaus.However, what none of the models can account for are the QUALITATIVE variables that affect a startup’s success.For example, in the case of Uber, factors like these:* legal frameworks of market;* perceptions of users, competitors and legislators;* people chemistry and dynamicsSo the 20X can be the catalytic effects of any of these qualitative factors and not just the pure quant numbers.This is the opportunity for economics — not just startup metrics. The qualifiers of success and not just the quantifiers.
Terrific post and thanks for churning this out.
Point well made, as it was in your LeWeb video in Paris – this was a very interesting point to pull threads on – Bubbles vs. huge valuations that can lead to sustainable companies. What are the best recent examples of companies that haven’t been able to sustain/justify the initial valuation? Groupon maybe?
Zynga is one that hits close to home
Color. Sequoia and Bain as investors in its seed round of $41 million and then…* http://mashable.com/2012/10…
For kicks, I went to Ycharts.com and looked at Twitter valuation. It’s compared to Facebook, Groupon, and LinkedIN. http://ycharts.com/companie…Then I did a Facebook chart with their price, EBITDA and Revenue for kicks to see what it looked like. You can also layer in a lot of other fundamentals.<img src=”http://media.ycharts.com/charts/efae03aeea2da184595ee3c371ba6ee6.png” alt=”FB Chart”/>FB data by YCharts
Want a better feel for this multiple? Start with a low multiple and crank it if you see a moat which assures you the the company will be around in 20 years, crank it up if the company has network effects, crank it up if the company has high switching costs, crank it up again if its margins are high and sustainable and scale (low marginal costs), crank it up again it the the customer base is not concentrated by a single event (due to regulation etc.) and crank it up if the size of the market is growing faster than the revenue growth, now look at the big lever, revenue growth, and youll know if that growth will be there in the future.
Ya, look at a company like $CME. Massive margins, huge moat and a regulated monopoly for all practical purposes.
Yep, compare to NYSE
purely regulation. SEC regulates in a different way than CFTC, which penalizes stock exchanges (in favor of big banks)
Nice summary of variables to work with. Think you missed any important ones?
Three letters: PEG
Aren’t high revenue multiples at such an early stage also due to the limited availability of a company’s shares? If Uber’s shares were public, the price might come down to earth — or at least to a realistic value based on the opinion of millions of individuals. But because the shares are only available to a select group of investors, Uber can demand a premium price. E.g if you want to get in on it you gotta pay a little extra for it.
self fulfilling prophesy?
To your point Fred, there are a lot of ifs. If that’s true, why not apply an expected probability then an expected value. If you’re expecting $40b is the normalized valuation based on what might happen in 3-5 years, and maybe this great company has 80% probability of achieving $40b of normalized value, why isn’t the value $32b? And since there’s a cost of equity/risk, why isn’t that discounted by [15%] per year into today’s dollars?Seems like the ibank is tricking people into paying full boat pricing as if 1) $40b today is worth the same as $40b in 3-5 years, and b) there is a 100% probability they will meet those projections.A few companies have done that, but shouldn’t the new investors get the pickup in value for taking that risk, instead of the company getting 100% of that benefit today?Or is it just a cool baseball card that’s worth what a fan is wiling to pay for it, and the rest is creative justification?
It is the price the market is willing to pay to invest in what every one perceives as a rocket ship. It is very possible that it may not turn out as people expect. There is no 100% probability of anything.
We should have a “cliche Friday”.
The rocketship analogy is powerful here. Uber has very Challenger-esque qualities to it. Fred has written before about product cycle and it is a guaranteed fact that disruptors are disrupted with increasing speed.
If that’s true, why wouldn’t one wait to buy the stock after 1-2 years…after seeing how close they’re tracking to 100%?$40b is slightly less than the mrkt cap of GM, which cranks out $8B of EBITDA.I get Tesla,Google and FBook. $40b for skimming taxi rides doesn’t make sense to me. Especially given reckless mgmt and regulatory risk…
I would not dismiss it as skimming taxi rides. Transportation is a huge market worldwide.Let us also understand that valuation is not an absolute value of anything. It is what the market is willing to pay at a point in time, based on projections and speculation on the future. It is also a function of perceived risk, available proof / data and the competition for the deal.Investing at $40 B is no guarantee that you will make or lose money. The future unfolds on its own terms 🙂
I respect that answer, appreciate it. I’m definitely not a buyer of its IPO, but then again I’ve missed many rocket ships (missing PCLN still haunts me) ; )
thank you. I have missed a few rocket ships as well. The thing with rocket ships is that they are hard to recognize in the early days. It takes a lot of conviction (and even clairvoyance) to see the potential of what something can be vs. what it is now. By the time they start to look like rocket-ships and everyone can see the form and shape, it is often too late.So, what is required to see a rocket-ship before it becomes one ?Often, the biggest hurdle is hubris and the ego’s presumption of knowledge vs. learning to listen to the market, question your own assumptions, and evaluate opportunities with a very open and even blank mind. This is easier said than done as most people tend to look for external validation and other people’s opinions vs. seeking to understand fundamental customer needs, motivations and behavior.The more open we are to everything we see and hear, and the more curious we become, higher the chances to connect the dots, sense impending market transitions and figure out who may be able to exploit these. I really believe that being non-conformist and right (on a consistent basis) requires less arrogance, more reflection, and more empathy.
See my comment below about Mundell-Fleming ISLM.Sure, the investment banks could slice+dice the model with Present Values using different mixes of discounted and interest rates.However, there are also INTANGIBLES such as brand recognition and even the qualitative dynamics of Uber’s mgmt team which goes into the Uber valuation — not just revenues and the quantifiables alone.——I am an ex-ibanker.
so doesnt that mean the intangibles would make it worth 42billion later
The “inflation” — or what bankers would call “a premium” — of the intangible value depends on Travis & Co being able to do this:* https://hbr.org/2010/02/fou…Brand valuation has different approaches.Take for example, Coca-Cola. Suppose every year for a decade they spend X on marketing and sales, of which X(b) is the amount they specifically invest in branding — e.g., logo design, registering the trademark, defending against mark infringement etc. X(b) is subject to an amortization schedule because the brand as an intangible asset is treated in accounting terms similar to tangible assets in this respect.Typically, a relative brand valuation = capital invested by brand X (EV/capital invested by brand — EV/capital invested by nearest competitor).where EV is the market equity value of the entire company.
OK, so what happens when the next round comes in and values it at an even higher multiple? Would that then restart the whole cycle?
i’m curious about this as well.
If the growth justifies a higher valuation, there are more sharks in the water and there is more demand. If the company continues to grow, the love fest can continue until and beyond a liquidation event. When growth tapers off contrary to expectations, or if the market senses new risks hitherto unknown, things tend to get messy. For companies still private, these can lead to existential crises, and a variety of misalignments among investors who came in at different times with different expectations.Another way to think about it is that valuation = function (proof + promise). Proof is real and based on actual performance and metrics. Promise is based on assumptions. With every passing day, as more unknowns come into the realm of the known, assumptions are validated or invalidated, and promise becomes less or more probable.When a super-high valuation is driven by 100% promise and no proof, bad outcomes are a very real possibility. Better place is an interesting example.http://www.fastcompany.com/…
great post.Growth is the source of all high orbit energy.High revenue or user growth drives premium valuation multiples. The reality of high growth + promise of future profits is enough to justify a super high valuation, even with zero or negative EBITDA, as long as growth continues to occur.When growth eventually declines, valuation returns to a normal orbit and traditional profit and revenue multiples are back in play.The competitive reality of the business ( i.e. the moats to dissuade new entrants, and maximize profits, and how long these can last) will need to be reflected in the revenue multiple. This may be the hardest to predict, but has a huge impact on future profits.And, if there are other adjacent markets that could be served in the future using the same technology, these provide additional oomph to the multiple. In Uber’s case, it is the car ownership market and the larger logistics play for other products.Post product market fit, growth is the only reality and truth that matters. Everything else – good and bad – flows from growth or the lack thereof.
Exactly. You should have written my post this morning!!!
The most unpredictable and biggest disruption is traditionally whenever a new business model comes into play – and is executed well. FB may look in free flight growth at the moment – though there are many first principles which they don’t or won’t ever include in their model which could turn them in the other direction rather quickly.
Great post.People actually throw the baby out with the bath water on these topics (I.e. They think all CEO comp is ridicous when a small # deserve every dime & they think the markets are drunk when Uber, FB & GOOG are actually outlandish businesses that deserve outlandish multiples ).
This is a fairly straightforward example of a “mathematical outcome” exercise when you project a ratio forward (which is not to say it is trivial, it points to the importance of ‘running the numbers’. After learning this lesson painfully earlier in my career, I used to often tell my team to “run the numbers” even for simple 10-minute exercise like this. Less to prove the point which was sometimes intuitive, and more to clarify the assumptions).The key assumption here is starting EBIDTA (non-negative/breakeven) and the EBITDA growth. The rest is a mathematical outcome. Starting with a very high revenue multiple and also a high revenue growth, the revenue multiple has nowhere to go but down.I actually think that just like you mentioned that some people can say a 20X multiple is nuts, the opposite can also happen. Viz – the prospect of future revenue multiple compression in a fast growing business is automatically assumed and can be used to justify – ‘The current valuation is only 5X revenues’ while revenues are growing triple digits’. But if the burn rate is (a)- very high and (b) -foundational to the current revenue, then the higher growth is simply not going to translate into the Opinc growth and even the 5X starting multiple may not be justified (while the 20X multiple just may work out in a different case as shown in this example). “Don’t look at cost of customer acquisition cost, look at customer Lifetime Value” — can be a very slippery slope. Thanks.
It is possible to know the unit economics but assumption of EBITDA is very tricky. In fact in most cases, it will be negative for the first few years as you are investing aggressively into the business. Also, CAC changes substantially over time with more success and a stronger brand.My point is that there is a lot more to estimating valuation than projecting a mathematical outcome. In addition to working out the viability of the model, it calls for judgements on size of market, adoption over time, competitive moats, ability to execute etc.I believe that understanding the real pulse of market behavior and the transitions under way are more fundamental to investment decisions. With partial data around adoption and growth, we need to fill out the rest by connecting the dots in our heads, correlating to old patterns we know and understanding the new patterns that may be forming.The ones who connect the dots early and call correctly make the big bucks. Which is why early stage investing is really more art vs. science.if there is anything the last decade has taught us, user adoption and growth matter more than everything else, including initial revenues and profits. You are likely to figure out the business model if users love the product and are asking for more.
Great post! A lot of people see high valuations and run scared away from an investment. Investors are trying to predict the future. A high current PE ratio (vs competitors) is a product of people driving up the valuations, with thoughts that this company has better growth potential vs it’s competitors. Earnings act a trailing indicator here where as people are trying to price out the future- which can lead to a disconnect. It’s hard for some people to swallow these high valuations, but if we understand why they exist they may not be so scary!
There’s a key point in the video which you made subtly at the end of this post, and that is to note that “only the very best companies attain those multiples, and in that magical moment of going public.”The key nugget is that “Stock price can continue to go up, as long as the company continues to grow revenues faster than the valuation multiples go down.”So it really boils down to the relationship between Revenue Multiples and Valuation Multiples.
Am I missing something about Uber revenue? Isn’t 1B their GMV? Their revenue is roughly their 20% cut of 1B? If so doesn’t that pin their value at 200x their revenue?
in the leweb chat i referenced this from Business Insider http://www.businessinsider….
Thank you for the link to the BI article. That makes more sense. Should have known you had your info correct.Thanks Fred! Love the email. First thing I read every morning.
Reading comments and a lot of people are mentioning Uber as a overblown start-up, is it just me or do people realize that the beauty of Uber is in their network. If automated driving cars become reality combine that with the Uber network and you get a Bonanza!
If automated driving cars become reality combine that with the Uber networkYou are talking about something that is so well into the future that for all practical purposes it should not even be considered.By the way for everyone who thinks there will be some kind of universal self driving cars all over just remember that at a certain time they thought everyone would have an airplane. It’s not the cost either. Planes could easily be affordable and cheap with mass production. And we are advanced enough now or even 20 years ago that that it could have happened. And there is definitely a need and demand. But it hasn’t. The reason is safety is involved and there is a high chance of fatality if things don’t go right. While I am definitely seeing computer assisted driving I am totally never seeing self driving cars the way people are talking about them in our lifetime. (Now of course if you build roads just for self driving cars then that’s another situation.)
That is a bunch of “ifs”To me business is all about the idea and how it’s executed (duh). So in the end if the idea is good (“low hanging fruit of opportunity”) things tend to work out as long as the people in charge, as well as the people that work there, aren’t total fuck ups. Also luck and all of that. And if the idea is bad things tend to turn out bad despite herculean efforts by the best and the brightest. It’s really that simple. All the financial engineering and comparisons are a good distraction if someone doesn’t want to understand the basics of the business. And whether it actually stands a chance of working out.This kind of dovetails with my theory on buying things. Instead of focusing on only buying something if it is “a good deal” focus on whether the thing you are buying you really should be buying at all.  The mistake people often make is deciding “I will buy if the price is $x but not $x + 5%.” The fact that people buy when there is a good deal is what makes people buy things that they don’t need. Dangle a good deal and people forget if they need what they are buying. Near universal human nature.Which brings us back to Uber. If Uber works out the multiple won’t matter. And if Uber doesn’t work out a smaller multiple won’t matter. Makes total sense. For example if you had bought real estate in NYC in the 70’s it would have been near impossible to overpay for that real estate. But yet many people (who were investors) probably passed on buying a property because they couldn’t buy it on the cheap. Someone who didn’t think like that (and might have been stupid) ends up the one with more money because they aren’t buying based on value. Exclude people who simply don’t have enough money to take a gamble.
Not saying this is wrong but my finance prof at Stanford GSB did the same calcs on the eve of the first Internet bubble bursting. I believe the example was pets.com. Yes, any valuation is justified as long as the underlying asset can grow — indefinitely.
Seriously though, who uses revenue multiples for valuation? They’re just an output of valuation using profitability multiples or other multiples (users, etc.). Haven’t ever seen anyone actually make an investment decision based off of revenue multiples.
Why do SaaS companies almost always trade at a multiple of revenues then? Why is this relevant and not just an output for them and not others?
The example here is an unusual one of a suddenly very profitable business that is also growing very fast- you don’t need that level of EBITDA. Salesforce has +/- zero EBITDA and a 7-8X sales multiple (of course it’s not worth as much as a taxi company…)- because it still has that top line growth.Companies with positive unit economics (aka customer lifetime value) should, in principle, spend as much as they can on acquiring as many of those users as they can- that will drive value, and also losses, as per customer revenue lags behind cost of customer acquisition, on an aggregate basis. The math of what that looks like on a revenue multiple along the way is “just” conference fodder (so agree with the “moment in time” comment, CRM shows you can sustain something a little less for a long time)The simple model for Uber makes the passengers are the “units” of economics here- not sure we have the data here to understand the dynamics/model this properly (?)
Just noticed that Lending Club is now trading at ~100x revenue according to this:http://www.bbc.com/news/bus…I guess their “multiple compression” will have to be dramatic in the coming years.
Thanks for the math Fred.Two of my stocks, Mastercard (value 100B) and Catepillar (Value 56B), each make over 5B per year profit before taxes. They are proven concerns with wide moats that have been around awhile.In comparison, no way an Uber, for instance, can be worth 40B with all the what if’s involved. But it doesn’t matter.All that matters is that at IPO the well-known and recognized company is sold to the public for more. Zynga, Groupon, GoPro (still doing fine), etc.And the differences in investors, in estimates of the future, in companies, in expectations, is what makes investing so damn fun.
That actually is helpful, but some of these numbers are determined by the risk free rate, which there are still questions about.
What’s the risk free rate? Factors that determine likelihood they’ll be knocked out by a competitor?
http://en.wikipedia.org/wik…The issue is that there may have been negative inflation rates/0 risk free rate (since usually the comparison is Us treasuries)
Can I just add that valuation is ultimately a function of CashFlow and that profit is just a ballpark made up figure?
no you cannot
Earnings are used for the P/E ratio, true, but the real health test of a company it’s the CF..profits include non-cash items for instance and therefore are inaccurate.
I would say the real health test of a company are more the qualitative values – CF may just be a signal of that success.
for example? ..i agree with you but I’m just curious on your thoughts..
Well, you could say the quality of relations within the company, how much do people actually want to be working there – how passionate are they?How well is the company being guided? How compassionate?Does the company use aggressive or bully-like tactics, or passive but present tactics?How a company is managed now will determine how it will succeed in the future, assuming the market is there and overall execution is done well.
Great post Fred. If we’re saying a 20x multiple is generally a high watermark for a public valuation, how does this translate to predicting the fate of privately-valued companies with valuations many times these kinds of multiples? Certainly there will be exceptions (social networks with little revenue but massive scale, ie. Instagram, others) but what about the others? If they can’t find an acquirer along their journey, is it too crude to say most of these co’s are screwed?
Buying a hyped pre-EBITDA company at 20x revenue for a 10% yoy returns over 5 years … seems a lot of risk for the return.
That’s exactly right
Thanks, Willy and Fred. Really appreciate the kind words and feedback.