Why are SaaS companies expensive to invest in and why will they always be?
The act of comparing is common in all aspects of our society. We believe that Usain Bolt is currently the fastest man on earth because the rules and objectives by which we identify and compare runners are clear (runners start at the same time, run the same distance and all are measured by the same system).
Yet, things get more intricate when you compare historical results. On a relative basis, who was faster, Usain Bolt or Donald Lippincott? The evolution of the sport has been tremendous in the last 100 or so years since Lippincott ran for an impressive 10.60 in the 1912 Olympic games. For instance, starting blocks were first used in the 1948 Olympics (although studies are inconclusive whether it improves performance of athletes). The 1940s also saw the introduction of synthetic hard surface that are now widely used in tracks that dramatically improve the composition and the elasticity of the tracks, enabling athletes to bounce faster and smoother. Lastly, progress in nutrition, training, and even in sleep have all improved how runners plan, train and compete.
Comparisons in the investing world
In the investing world comparisons are also common. For companies planning to become public, the lead investment bankers hired in preparation of the IPO create comparable analysis (or ‘comp tables’ or only ‘comps’) to be used throughout the process. By nature, a comp table is a list of relative valuation metrics of competitors or players in the same geography and of similar or adjacent industries and sizes and it is one of the many valuation metrics applied before pricing the shares of a company.
Comps are also part of every M&A transaction and venture-capital fundraising round for the need of establishing patterns. At the end of the day, these analyses are useful tools to mitigate risks and anchor expectations of the people involved in these transactions (“are we under/overpaying for this acquisition relative to other similar deals?”)
For instance, when Salesforce acquired Tableau in 2019 the company hired Goldman Sachs to analyze the transaction and, as part of the engagement, the bank compiled a list of public-traded comparable companies to define the fair value of Tableau.
This week, the team at SaaS Capital have published a revised version of their SaaS Capital Index and Bessemer Venture Partners revamped their Nasdaq Emerging Cloud Index website, both of which are highly recommended sources of information related to comparison analysis of Software-as-a-Service (SaaS) companies and inspired me to summarize some valuation metrics data around this valuable data sets.
The SaaS Capital Index is made of 49 SaaS companies that were public traded at a given point in time between May/08 (Autodesk IPO) and Dec/19. Over the years, the SaaS IPO market have gained significant interest from the general public given their high-flying valuations for business models that were — in several cases — unclear for many. When looking at the median Annual Recurring Revenue (ARR) multiple over this period it increased from 3.5x in the end of 2009 to 8.8x in 2015 to 10x by the end of 2019.
But behind this upward-trending slope with a relative low variation (standard deviation of 2.5 and median of 7) there is a lot of variability in the sample as we can see in the figure below, where I added the difference from the highest to the lowest ARR multiples for each month. From 2008 to 2012, investors paid no more than 15x on an ARR multiple basis for SaaS stocks. One reason can be attributed to spillovers from the 2008–09 financial crisis in which investors moved to safer assets.
Others might claim that the shift to the cloud was still nascent and full of uncertainties in the development of its business models and adoption of enterprise customers.
What happened in 2012–13 changed how the markets perceive SaaS Companies
The trigger event that transformed investors’ risk-taking appetite to evaluating SaaS companies with higher and higher ARR multiples was the Splunk IPO. While many ‘experts’ called it reminiscent of the dot.com bubble era, Splunk solved a complex problem and was growing faster than its peers (speaking of comparisons: Splunk current valuation is 7x larger than its IPO price).
The year of 2012 was also the year that many other high-growth companies became public, such as Workday, ServiceNow and Palo Alto Networks. In fact, 2012 was the 3rd most prolific year for SaaS IPOs since 2008.
Up until the 2012–2013 IPO class, SaaS companies would post revenue growth from mid-10%s to mid-30%s (Salesforce grew 36% in 2012, an average of 31% from 2010 to 2019 and never above 40%).
What Splunk, Workday, ServiceNow and Palo Alto Networks introduced at unprecedented scale for public markets investors was the benefits that network effects bring to successful companies, where top performers capture a substantial large part of a market — and therefore earn outsize returns with high barriers to entry — and the remaining competitors are left with very little and will typically compete in price. These four companies alone posted average revenue growth of 76% in 2012 and as high as 99% in the case of Workday.
As a result, investors flocked to them and became ‘comfortable’ in paying 30x or more in ARR multiple. For instance, Workday (WDAY) traded for an average of 28x ARR during 2013 and as high as 33x in September of that year. In the early days of 2014, the market found a hot company in FireEye after the $1billion dollar cash-and-stock acquisition of Mandiant announced on January 2, when shares of FEYE jumped in excess of 20% and remained elevated until its quarterly results.
The ‘stable’ period of 2015–2018 and the widening patterns of 2019
Then, from 2015 to 2018 the difference between the highest ARR multiple and the median cut to mid-to-high single digits and only briefly touched 10x or more twice (once in mid 2015 and another in late 2018). Again, one argument is driven by the fundamental of the companies: with the slowing U.S. economy in 2015 and 2016 and the ‘uncertainties’ related to the 2016 election cycle only one company posted +50% in ARR (Coupa). Another is market-driven, with the median ARR multiple rising from 6.6x in December of 2016 to 11.4x in September of 2018 with the Tax Cuts and Jobs Act of 2017 playing an outsized role in boosting the economy and the broad stock markets.
The start of 2019 saw a widening in the relative valuation of high-flying SaaS companies (growing at ARR +50%) compared to the others as investors seek for ‘best-of-breed’ in selected markets until the events that are referred to as the ‘Rotation From Growth-to-Value’ in the summer of 2019 crushed the valuation of the most expensive SaaS companies after investors moved to safer assets following the inversion of the yield curve.
The following chart seems a mess at first (and it continues to be even after looking at it for a while), but it is a stab in having a comparable analysis with a different twist. For this graph, I grouped companies segmented by ARR growth by color: the highest growth SaaS companies are displayed in blue, followed by a cluster in green-ish and lastly the ones in orange.
The 2017 to 2018 period saw a barely indistinguishable market on a relative ARR multiple basis for the SaaS companies (The outlier was Veeva Systems for its strong growth in the healthcare industry). On average, Investors payed the same multiples for ‘underperforming’ SaaS companies relative to the high-growth ones until the beginning of 2019, when there was a rally of the highest-flying stocks including OKTA, VEEV, COUP and CRWD.
In the ‘Rotation From Growth-to-Value’ event (approximately July to October of 2019), it was clear that investors discounted mostly the highest multiple stocks. Alteryx saw a 7.7x decrease in its multiple during the period, followed by 7x in Okta and 5.3x in Anaplan. On the other hand, Workday’s multiple dropped by 3.3x, HubSpot 1.8x and Salesforce actually increased by 0.66x.
By the end of 2019, this was the picture of Growth Rates x Valuation Multiples. The correlation is positive between these two variables and growth rates can explain about 30% of the valuation premium for a given company of this cohort.
Interestingly, when I plot the number of SaaS IPOs against U.S. GDP growth, it can be observed a positive correlation between the variables. If we ignore the 2008–09 years of crisis, the R2 improves from 0.13 to 0.21, so while U.S. GDP should not explain the number of SaaS IPOs, it can be symptomatic because an IPO can be as much of an individual story as it is about how similar companies perform in the overall economy.
In summary, SaaS companies are expensive on a relative revenue basis — but they remain consistently expensive through long periods of time as long as their business models continue to deliver. As discussed earlier, there is a positive correlation between growth rates and ARR multiple and the market will continue to call for imminent bubble bursts even if we know how bad we are at predicting the future.
As Bill Gates said in 1996:
“We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten”