A DEEPER LOOK INTO SNAP’S IPO
So Snap (AKA SnapChat) issued their S1 filing for an IPO last week. This is big news! One of the more known Unicorns is going to brave the waters of the public market.
Make no mistake, from the perspective of the company, this is just another fundraising round, but this time under the watchful eye and relentless scrutiny of the SEC and every single tech-finance journalist in the world.
From the perspective of the early private investors, this is a potential windfall — a financial harvesting event that has been years in the making.
Many have already written about the Snap IPO, looking at the revealing numbers in different ways. You can read what Pitchbook, CB Insights, and Marketwatch all have to say, but I wanted to take it a step further and look at the (few) numbers that have been recently published in a different light.
Thoughts always welcome.
Most of the articles you will read on the Snap IPO compare the company to other mass-market social networks, like Facebook or Twitter, looking at metrics like price per user. I want to compare Snap to itself (gasp!) using their (minimal) available data.
Snap first incorporated ads as a revenue driver in Q1 2015, when they reached 80m daily active users. Since then, ad revenue has increased substantially. More importantly, ad revenue per user has increased. This is a good sign. It means that Snap’s growing user base is becoming more valuable over time, as should be with networks.
But let’s take this a step further and look at the cost component. In general, as a company matures, we want to see that revenue per user (or customer), eventually cross the line of operating cost per user (or customer). This is generally considered the break-even point, where a company is able to reach operational profitability month over month (or quarter over quarter). In other words, startups typically have much higher costs/unit than rev/unit, but in the long run these two need to converge, otherwise the company is not financially sustainable.
In the case of Snap, cost/unit and rev/unit (measured in DAU) were converging slowly until Q2 2016, and then costs/unit spiked. Now, this is not necessarily a bad thing. Snap pumped a lot of money in to R&D and COGS over the last half of 2016, and these additional expenses could very well be justified. (Spectacles maybe?) So while this is interesting, it is certainly not a red flag. But over the long run, say the next five-ten years, we would like to see these two lines converge, and not drift farther and farther apart.
A third interesting parameter that we can look in to is CAC vs user growth. Given the little amount of detailed information in the S1 filing, we have to create a rough proxy for CAC. This will obviously not be exact, but it will give us a sense of the trend in CAC. Specifically, we look at Sales and Marketing expenses per quarter, divided by the number of new DAUs at the end of the quarter.
The logic behind this is that we assume that S&M outlays over the course of a quarter are dedicated to increasing the amount of DAUs, measured over the same time period.
When we do this, something seems a little concerning:
What we see is that in the last two quarters of 2016 Snap spent quite a lot of money on customer acquisition costs, but failed to deliver new DAUs. CAC trended up, but DAU growth trended down.
Of course, we don’t know for sure because this is the first time we are seeing anything internal from the notoriously secretive company. It could be that this effect is completely justified. But in general, this is quite concerning and raises just more questions. What portion of the CAC went to the spectacles, for example. What about LTV? (To get LTV we would need churn, and the word “churn” does not appear at all in the S1 filing).
For a company whose primary current source of income is ad revenue from a massive user base, increasing that user base is the MO of Snap’s entire existence. Therefore, such an aggressive increase in CAC without subsequent new users could pose an existential problem.
Of course, the private investors will already be cashed out and richer by that point.
All of this reminds me of two recent academic papers (“The Rise of the Unicorns — How Media Affects Startup Valuations” and “Valuation of IPOs with Negative Earnings”) by Severin Zorgiebel from Goethe U. in Frankfurt.
Zorgiebel points out that quite a lot of companies IPO with negative earnings (over 40%), but what differentiates VC backed IPOs is the media blitz, which is unique to startups. In his words:
“IPOs with negative returns might be different and more exposed to marketing hype than other IPOs. This effect is fueled by the influence of VCs and underwriters… These high valuation levels are the result of marketing campaigns performed by VCs and underwriters. These results hold in a variety of robustness tests. The marketing of IPOs influences the investors’ perceptions of these companies, thereby increasing demand due to high growth expectations and generating high valuation levels. These valuations do not appear to be based on financial fundamentals and might be indicators for overreactions.”
So Snap will most likely raise its $3B, VCs will most likely make money, and regular public market stockholders will take on the risk of poor financial returns. The circle of life will in the startup ecosystem will continue.
