After reading Alex Wilhelm’s sixth entry in his “The Changing Value Of ARR” series — an awesome way to stay up to date on startup valuations! — I typed up some thoughts on when revenue-based multiples should and shouldn’t be used as valuation tools.
Companies are only ever worth the cash they’ll generate in the future (discounted back to today’s dollars). Even if you don’t (or can’t) do a full discounted cash flow (DCF) analysis, and just use a simple multiple to value a company, you’re just taking a shortcut to the same result. Every multiple-based valuation can be converted into “implied” assumptions about a company’s future cash flow generation (such as growth and margins).
Traditional profit-based multiples, like P/E (a.k.a. price-to-earnings or Market Cap-to-Net Income), work because “earnings” (whether Net Income, EBITDA, etc.) is a pretty close approximation of cash generated in any period. So you can fairly compare a P/E multiple from a low-margin company like Overstock.com to that of a high-margin company like Oracle, and not worry about the difference (the same goes for industry-level comparisons — the oil & gas sector to homebuilders, or emerging market stocks to US stocks, whatever). The different margins are already reflected in the denominators of the two multiples, so you don’t care how much revenue it took each company to get there. The biggest remaining difference between two companies or sectors is whether that “earnings” denominator is expected to grow or shrink!
Revenue-based multiples like EV/Revenue, however, are pre-margin! So you end up with revenue numbers of widely mixed quality. For every dollar of revenue, Overstock is scraping out less than a penny of profits, while Oracle generates more than 30 cents. Revenue is a pretty ugly approximation of cash flow generation.
As a result, comparing revenue-based multiples from companies with different margin profiles is almost entirely meaningless. And, for growing, early-stage companies in particular, gross margin profiles are what we care about. The preference for gross margins is because there’s an assumption that operating expenses (like engineering and overhead) are “fixed” costs that’ll get dwarfed by revenue as the company grows, while costs of goods are “variable,” and will continue to grow roughly in line with revenue forever. Nobody cares about Uber’s OpEx today — the cost of the H/R department, the cost of Engineering, etc. Those will eventually be tiny compared to total revenue. But, I guarantee investors are worried about the amount Uber pays its drivers, because the number of drivers and rides will always grow roughly in line with revenue.
So, just like with profit-based multiples, future growth is one of the most important drivers of revenue multiples — as Tomasz Tunguz (Tom Tunguz) and Alex Wilhelm clearly illustrate, there’s a strong connection there for SaaS companies! But, unlike with profit multiples, there’s a second key factor that affects the relationship between a company’s revenue and its future cash generation — margins. The simple growth-to-multiple analysis that works so well for profit-based multiples should become a three dimensional growth-and-margin-to-multiple analysis for revenue-based multiples.
Box vs. Salesforce
So why is the correlation that Mr. Tungusz finds still so strong? Because most SaaS companies have pretty similar margins — it’s just the nature of software to be in the 70–90% gross margin range, so you can generally build those charts and ignore the margin factor altogether…
But, that may not always be true. Take Alex’s example — while today both Box and Salesforce have similar gross margins (both in the 70%’s) there’s a big difference in future prospects. While Box is growing its revenue rapidly, it’s gross margin is falling. Prices across Box’s industry are dropping, as Google, Apple, Microsoft, et. al. try to compete in the enterprise by giving away storage for free. Hosting services are becoming more commoditized, and the amount of profit Box (and Dropbox, etc.) are able to collect from a dollar of revenue is slowly shrinking. Going from just over 78% gross margin in FY’14, to 70% gross margin over the last twelve months and just 69% in the last quarter is a scary sign. Salesforce’s gross margins, meanwhile, are flat — there’s less competitive pressure on the price their product sells for.
In the long run, the margin profiles of these two businesses may diverge, and as a result, revenue-based multiples will become less and less comparable. In fact, it may be the case that Box’s lower EV/revenue multiple is the result of investors already building structurally different gross margins and cash flow margins into their long-term DCF analyses.
