In a previous post, I wrote about why investors in SaaS companies feel forced to use EV/Revenue to value them. With profits absent and a subscription revenue dynamic that makes near term profitability a poor indicator of long term success, it is little surprise that investors have resorted to putting a multiple on a metric all the way at the top of the income statement. Below, I'm going to break down some challenges of this approach and potential solutions. There are two main issues I'll cover: accuracy and delegitimization.
In general, the further down the income statement it rests, the less valuable a metric is for giving insight into future cash flow generation. Consider two tech companies with $100m of revenue each, each growing 30%. One is a SaaS company with 80% gross margins, where COGS consist of AWS costs and customer support. The other is a vertically integrated online retailer in a commodity category with 15% gross margins. It goes almost without saying that the long term margin potential of the SaaS company is far higher, and that it is worth far more despite a similar revenue level and growth rate. But from a purely EV/Rev multiple lens, these companies look very similar. In fact, even if we try to adjust for margins today, we get tripped up: both companies might be running at a -5% FCF margin. Non-gross profit generative revenue dilutes both profits and losses- it both lowers long term margins and can make near term losses seem less extreme than they are relative to the long term earnings potential of the business (lessons from Groupon!). If we rely on the financial metrics alone, it is easy to think that the EV/Rev multiples these companies deserve are in the same neighborhood: they are growing the same pace and both losing modest amounts of money after all, but in reality that's far from true. For a commodity retailer, a 5% net income margin in the long term is nothing to sneeze at, whereas for a SaaS company with 80% gross margins it isn't hard to see 30% net income margins one day in a low growth scenario when the sales floor is gathering dust and a private equity company owns it. A company earning a 30% net income margin valued at 20 times earnings (reasonable for a high quality business) will be worth 6x EV/revenue, whereas a company at a 5% net income margin at 20 times earnings will be valued at just 1x EV/revenue. Even as high growth start-ups, SaaS companies and low-margin retailers clearly deserve wildly different EV/revenue multiples.
That's not a hard one to follow, but what gets tricky is when two companies nominally in the same industry have business lines that create meaningful differences in their gross margin profiles. Here are three common examples I see in software regularly:
1) Revenue is recorded gross but is collected net (a common practice for ad tech companies)
2) A company has a large services business running at low to negative gross margins
3) A company has a payments business as part of its software that includes per transaction COGS of some kind
Here's a quick map of various public software companies (mostly SaaS) by gross margin. None are sitting at the retail level per se, but the differences are plenty big enough to make a difference in the long term margin profile and current profitability optics of the businesses as discussed in the earlier example. It is far from a cure-all, but this is why I generally use EV/GP as my go-to valuation metric for tech companies, whether they are e-commerce retailers, ad-tech DSPs or SaaS companies.
I'm also of the opinion that EV/Rev valuation multiples do a disservice to SaaS companies in the long run, because they perpetuate the dual ideas that they are financially rickety constructs susceptible to market downturn and most investors assume that they will decline over time as companies growth. Both are wrong- I won't spend much time on the financial side because it relates to well understood themes that I spoke about in my last post regarding subscription revenue models, but suffice to say SaaS companies are high quality businesses that in the long run should not be too exposed to business cycles. On the subject of multiple compression over time, things get interesting. Investors often assume that because only unprofitable companies are valued on an EV/Revenue basis, any EV/Revenue multiple above 4 is "high" and will inevitably compress with time as growth slows. Reality suggests this intuition is wrong- in fact, striking just one outlier (SHOP) from the scatterplot below is enough to make the statistical significance of growth as a driving factor behind valuation on its own disappear. Instead, we see slow growth companies like Aspen Technologies, Ansys and Blackbaud trading at valuations well above many higher growth names. What gives? Investors care about profitability too, as it happens, and so long as companies make progress on profitability as growth slows, there is no real reason why the multiples on all but the most expensive SaaS companies need to compress very far over time:
One final way of considering this counter-intuitive result is using the SaaSco example I've referred to in the past. SaaSco is a fictitious, moderate-quality SaaS company I have built a model for where I have perfect visibility into important metrics like CAC and churn. In an earlier post, I showed how SaaSco is profitable only below 15% revenue growth. It can also give us interesting insights on valuation. By placing a FCF multiple on it in old age (<5% revenue growth) to get a relatively accurate valuation, then discounting back at a standard discount rate, we can get an idea of how the valuation should have trended on a EV/GP basis over the life cycle of the company. The result is a nice fit with the scatterplot above: over time, the proper EV/gross profit multiple for a well functioning SaaS company initially compresses as growth slows, but then levels off and can even increase in old age as profitability becomes a key driver of valuation.
To grapple with this challenge of using a simple, often single digit multiple to accurately capture a multi-decade growth story, there are a few approaches I've found useful. One is to forecast far in the future, and then apply a steady state margin profile and discount back. For the public SaaS companies I follow today, I generally look out to 2025, estimate how large they will be and then "steady-state" them by applying a low-growth margin profile tailor-built for each company to the given revenue level. Another approach is applying that same "steady-state" margin profile to today's revenue and valuing the company on "earnings power" rather than revenue. In both cases, the key is that incorporated into the approach is some semblance of the long term margin profile of the company- in the end, it is the easy to fall into myth that all dollars of revenue are equal in their profit generation potential that makes EV/Rev such a weak and potentially misleading approach to valuing companies.