ARRG: How One Neat Startup Valuation Tool Can Fall Short

What is a startup worth? There are as many ways to answer that question as there are startup valuations, and we have indulged in its pursuit on these pages quite often.

To pick one example, back in February, Crunchbase News spent some time noodling on relative startup valuations using a nifty tool called annual recurring revenue to growth (ARRG). An invention of Bessemer, a venture shop, the metric is an attempt to understand why startups are worth as much as they are compared to their public comps.¹

Today we’re going to take a second look the ARRG startup metric. Our goal is to better unpack its substance, flag what we missed the last time we employed it, and give ourselves a better understanding of how investors value quickly growing private companies.

Some Loose Definitions

Let’s start with some definitions and explanations. We’ll start with annual recurring revenue, better known as ARR. ARR, per Zuora’s Tien Tzuo, is “the amount of revenue that you expect your subscribers to pay you every year.” There are more technical definitions out there (SaaSoptics has a good one here) but Tzuo’s works. (More on Tzuo’s latest here.)

In practice, startups sometimes take their last month’s recurring revenue and multiply it by twelve to calculate their ARR. So if your company had $100,000 in recurring revenue last month, your current ARR would work out to $1.2 million. (More here on slide eight.)

Next, let’s talk about ARR multiples. Leaning on the simplest possible definition, a company’s ARR multiple is its valuation divided by its ARR. So a company worth $100 million with ARR of $10 million would have an ARR multiple of 10.

Some companies have higher ARR multiples, and some companies have lower ARR multiples. Investors are sometimes willing to pay more or less for recurring revenue depending on the company in question. Three factors contribute to how investors value recurring revenue: revenue quality, revenue cost, and revenue growth.

Revenue Quality

Revenue quality refers to the margin generated by ARR. If a company’s ARR has gross margins of 90 percent, that’s quite good. And valuable! A company with ARR gross margins of 70 percent would likely be worth less than the company with gross margins of 90 percent, all other things held equal. That’s because a company with higher gross margins will likely be more profitable in time than a company with slimmer gross margins.

Revenue Cost

Revenue cost refers to the acquisition cost of revenue. If a company has to spend $1 to generate $1 in annual recurring revenue, it is likely a better (and thus more valuable) company than a firm that has to spend $2 to generate $1 in annual recurring revenue. And a company that only has to spend $0.50 to generate $1 in annual recurring revenue is better than both of our other hypothetical firms.

Revenue Growth

The faster a company is growing the more investors are willing to pay for its revenue. That means the faster your little company is expanding its top line the greater its ARR multiple will likely be.

If your ARR grew 100 percent over the past year, that’s great, but if it grew 200 percent, that’s even better. And investors buying into your company today will, all other things held equal, be willing to pay more today for your current ARR. Faster growth now implies more ARR down the road, meaning that investors may be willing to overpay for today’s ARR, but they may be buying future ARR at a discount.

There is a tension between the three. Some companies will drive up their revenue costs to increase their revenue growth. Some companies will add new, lower quality revenue streams to drive revenue growth. But in the valuation game, comparisons between companies tend to focus less on revenue quality and revenue cost because investors are more interested in revenue growth. In less-mature companies, gross margins and customer acquisition costs can be nascent, changing, or hard to parse. Revenue, however, is reasonably clear.

And that’s why the relationship between a company’s ARR multiple and its growth pace is interesting. Their interplay impacts nearly every modern software company, startup or mature tech. This brings us to ARRG.

ARR, ARRG, Argh

Crunchbase News previously covered Bessemer’s ARRG metric. The metric discounts a company’s ARR multiple using its growth rate as a weight to help bring buoyant multiples back down to Earth.

In the current boom times, venture capitalists are paying lots for growth, which, per our prior discussion, means that they are shelling out more money for present-day ARR than they might have in the past. But ARRG can take some of the bite out of the historical comparison.

Per Bessemer, ARRG is a metric calculated in the following way:

What this means is that if your company is growing at, say, 200 percent, you can cut your current ARR multiple in half. This is an interesting way to show that not all ARR multiples are made equal. (And thus, if you are an investor, that you are perhaps not overpaying for private company shares with your LP’s money.)

Bessemer used this metric in its State of the Cloud Report 2018 to show that, while private ARR multiples are dramatically higher than public ARR multiples, private market ARRG multiples are close to public market ARR multiples. This implies that startup valuations for ARR-generating companies aren’t too bonkers.

The last time we reached this point here is what Crunchbase News wrote:

The gambit at play here is that Bessemer doesn’t show the public company ARRG line in the example. Only comparing private ARRG to public ARR is a bit non-GAAP for my tastes.

What might be fun would be to show the gap between private ARRG and public ARRG multiples.

And that’s where I should have done more work.

Over the weekend, self-described “SaaS enthusiast” Geoff Byron asked us how ARRG multiples work for companies that are growing at less than 100 percent per year. My thought, at first, was that the formula should hold and that you just put in a smaller denominator (growth) than you would if the firm was growth at more than 100 percent per year.

But that doesn’t work. If you put in a number smaller than 100 percent into the formula’s denominator, the company’s ARR multiple is stretched, not shrunk. So while the ARRG metric worked well for Bessemer’s effort to show that private software companies ARRG results were similar to public market ARR marks (as the startup crew under inspection was growing at far above 100 percent), it doesn’t work for companies growing less than 100 percent annually.

We can interpret the situation in one of two ways:

ARRG is a fun way to show that companies doubling or more each year are not too overpriced, and it is not a tool designed to interact with all private companies’ valuations.
ARRG is designed to reprice companies both growing more and less than 100 percent yearly, bolstering the companies that are doubling and better, and dinging companies that aren’t hitting the 100 percent year-over-year growth mark.

Either way, we left too much work on the table before. That was my mistake.

I hope this was at least somewhat helpful. My key takeaway is that investors will always talk up their book; however, when they do, we need to triple check the math as things can slip by. Shoutout to Geoff for the help.

As we have linked to and discussed before, it was once argued that perhaps startups ARR multiples should fall under a public comp’s multiple. After all, their shares are illiquid, making a discount responsible. That logic is solid, but it also won’t get you any damn allocation in 2018.