Are Startups Over-Valued?
To Answer This Question, Start With the Public Markets
Like many VCs, we are preparing our 2020 year-end results and letter to our investors (limited partners, or LPs). As has been the case for many VCs, our results have been good; in the wake of the pandemic and all the carnage that is happening in other parts of the economy, embarrassingly good. If history is a guide, the question I expect to hear most from my investors (limited partners or LPs) is: these results look great today, but are startups over-valued? Can this bull market for tech, in general, and tech startups, in particular, continue? It is a question a lot of investors are asking themselves.
To answer that question, we need to start at the top of the market and flow down the food chain. For the tech sector, two of the biggest market gorillas out there are Google and Amazon, each of whom conveniently announced their 2020 earnings today. Both companies, not unexpectedly, crushed it:
- Google generated $183 billion in revenue, $41 billion in operating income, and generated $65 billion in cash flow through their operating activities.
- Amazon generated $386 billion in revenue, $21 billion in operating income, and generated $66 billion in cash flow through their operating activities.
Google is now worth $1.3 trillion while Amazon is worth $1.7 trillion — each one a 2x increase from “peak pandemic panic” in March. What drove those valuation increases? Not sales growth: Google grew revenue by only 13% year over year while Amazon grew by 38%. Those growth rates are strong for such mature companies but don’t explain 100% valuation growth. How about cash flow growth? Google grew operating cash flow by 19%. Amazon grew operating cash flow by 71%. OK, that’s better, but still doesn’t fully justify a 100% growth in valuation.
Valuing mature companies like Amazon and Google, in theory, is a straightforward exercise (we teach it in the first semester of the first year at Harvard Business School!). The value of a company (similar to any asset) should be the current value of all future cash flows, discounted back to the present. So when the market values Google at roughly 20x their operating cash flow, they’re assuming that Google’s 2020 results can be replicated at least 20 more times — or at least the equivalent of that if you discount all of Google’s future cash flows to the present. I know that’s a gross oversimplification, but hang in there with me.
There are two wicked big assumptions that you make when you try to value a company: how big will their future earnings be and what is the discount rate that you use to discount future cash flows to the present. Future earnings are a guessing game based on past operating results and your belief about the future. The first “aha” about why valuations of tech stocks are soaring is that the market believes that the pandemic has improved the odds that tech companies like Amazon and Google are going to generate larger earnings and more free cash flow in the future.
Regarding the discount rate: what rate do you choose? The discount rate is based on an assumption about interest rates or cost of capital: if interest rates are high, then cash coming in 20 years is a lot less valuable than cash coming next year. If interest rates are low, then cash coming in future years looks a lot more valuable.
Because tech companies are all about the future and the promise of future cash flows, most valuation models of tech companies are very sensitive to changes in the interest rate. To see what’s been happening to interest rates, here is the 10-year treasury rate over the last 150 years:
Hard to miss that trend line! A year ago, the 10-year treasury rate was 1.9%. Today, it’s nearly half that figure. That means all the financial valuation models that everyone applies to future earnings have been recalibrated, making future cash flows more valuable. And that’s the second “aha” about tech valuations: lower interest rates are making the net present value of tech company cash flows even more valuable — simply because the discount rate applied to their future earnings has gone down.
Moving beyond Amazon and Google and looking at smaller public companies, you can see the impact of these two forces. I don’t want to analyze our portfolio company, MongoDB, whose valuation has increased 4x from the low days of March 2020 to a $24 billion market cap today. But a similar story can be seen across many other, similarly-sized software companies. Hubspot, for example, has seen a 3x increase in its valuation since March 2020 to $18 billion. Yet, non-GAAP operating income grew only 35% and revenue grew 32% in the first nine months of 2020 as compared to 2019. Strong growth, for sure, but does it justify a 3x valuation gain? Again, a more optimistic belief in future earnings growth and a sharp drop in interest rates, resulting in an increase in the value of those future earnings, makes the valuation model soar.
Public Valuations Drive Startup Valuations
So now that we have deconstructed and gotten to the bottom of the soaring public market valuations in tech, we can turn our attention to exploring why private startup valuations are rising so dramatically. Only a year ago, the rule of thumb was that fast-growing, industry-leading SaaS software companies were valued at roughly 10x revenue in the public markets. Even though they didn’t have any earnings, private startups were valued at a comparable level because VC investors used the public comps as a rule of thumb — if the core business model and growth potential were the same, it was natural to assume that the revenue multiple would be a good barometer since earnings would come with scale, as evidenced by the more mature companies described above.
Today, the public SaaS comps for the fast-growing, industry-leading companies are more like 20x revenue. Bill Gurley’s well-written post a number of years ago about the keys to the 10x revenue club is still on point in terms of characterizing the varying quality of different business models, but the title needs to be revised to “the keys to the 20x revenue club”. Why? Because of a macro belief that digitization has accelerated in a post-pandemic world (as Satya Nadella famously said, “we saw 2 years of digital transformation in 2 months”), which will lead to faster growth in future earnings. And at the same time, a drop in interest rates has meant those future earnings — which had been 10–20 years out and so highly sensitive to interest rate drops were less heavily discounted and so now more valuable.
If the public comps are 20x revenue — roughly 2x what they were a year or two ago — then the private comps follow. In our work at Flybridge, the new rule of thumb is that a $10 million revenue SaaS startup growing fast and well-positioned can raise additional capital at a $200 million valuation or 20x revenue. The valuations compress a bit when growth-stage companies get closer to becoming public companies — e.g., a $50 million revenue SaaS startup might achieve a 15x revenue multiple or $750 million in value — because the M&A air is thinner at that scale (i.e., fewer buyers) and private investors want to be rewarded for selecting those companies that can get through the grueling IPO process (less grueling now with SPACs, but that’s a whole ‘nother story). Looking at Pitchbook’s valuation trends data (the latest is their Q3 report), the sector-wide data supports my subjective market observation: valuation in later-stage deals grew 70% year over year.
Early Stage Valuations Not Affected by Increase in Public Market Valuations
If public market multiples are up 2x, affecting later stage private market multiples, how about early-stage market multiples? Here is where it gets really interesting. My subjective read of the market is that pre-seed and seed valuations — the end of the market where we invest in as a typical entry point into a startup — have not dramatically changed. At Flybridge, the prices we are paying for equity in pre-seed and seed companies just have not changed all that much. Yes, the occasional hot company at the Series A stage will command a valuation that looks more like a later stage one if they get credit for forward execution. In other words, an investor says, “I believe you’ll do exactly what you say over the next year, so I’ll give you full credit for this year’s plan and pay for future growth just so I can get into this opportunity.” But by and large, pre-seed stage valuations remain in the $5–7 million post-money valuation and seed-stage valuations remain in the $10–15 million post-money valuation range.
Again, a look at the Pitchbook data supports this observation. They report that the median angel (i.e., pre-seed) pre-money valuation in 2020 (again, through Q3) was $5 million (so call that $6–7 million post-money), and the median seed pre-money valuation was $7.5 million. Both figures flat in 2020 as compared to 2019.
Early Stage Valuations: Cap Table Math vs. Comparables Math
Why haven’t these valuations changed while the later stage valuations and public valuations have? I can speculate two reasons. First, the earlier part of the market is the more inefficient part and the supply of capital has not grown as rapidly in the later stage. Thus, prices remain stubborn. Second, the price of a seed-stage company is arbitrary. At that pre-revenue stage, they are all worth zero — nothing substantial has been built and not real value has been created yet. Therefore, valuations are more about cap table math, driven by rules of thumb regarding how much investors want to own (typically 15–20% in aggregate at that initial stage) as compared to the amount of initial capital that a founder wants to raise to run the set of initial startup experiments for 18–24 months (typically $2–3 million)? Because the cost of experiments continues to be low (and arguably has dropped even further recently thanks to the cloud, open-source, and no code/low code), the amount of capital raised in seed rounds tends to be similar year over year. Founders are quite good at focusing on executing the experiments that matter the most during that initial 18–24 months to hit their key valuation inflection points. And if the prevailing math is cap table math (i.e., ownership) versus comparables math (particularly when there is no revenue off of which to calculate any comparables), it is unlikely that this valuation trend will change.
Conclusion
All of these trends bode very well for early-stage investors. We get to invest at a relatively low (and only modestly rising) valuation in promising startups who can rapidly gain valuation momentum with traction and achieving product-market fit. At least until interest rates rise, and predicting that is way outside my wheelhouse. That’s the message we will be sharing with our LPs.
