Last Thursday, Snap, the maker of Snapchat, submitted its S-1 filing to the SEC, initiating its Initial Public Offering process. Snap’s impending IPO has been the subject of a great amount of commentary for the past several months, and in late 2016, the company began honing its messaging around the event: that the company is more akin to Facebook than Twitter and fits nicely into a duopoly narrative.
Almost immediately after the S-1 was made public, observations from analysts, financiers, and the general consumer technology punditry began circulating the internet — with responses to the revealed financials and user base growth numbers being overwhelmingly tepid, if not altogether jaundiced.
It’s easy to understand why. Snapchat’s quarterly user base growth has slowed significantly over the past year: from 14% between Q3 2015 to Q4 2015 down to just over 3% from Q3 2016 to Q4 2016.
Snap is also losing a massive amount of money: in Q4 2016, the company lost nearly $170MM on revenues of $165MM.
But perhaps more worryingly, topline revenue growth is slowing along with average DAU levels: revenue only grew by about 30% from Q3 2016 to Q4 2016.
Snapchat’s slowing revenue and DAU growth rates are the most obvious concerns for potential investors, and they are the primary concerns cited in most dissections of Snap’s S-1 (probably the most comprehensive of which is this Quora answer by Terrence Yang). But I believe that two other risks faced by Snap which have mostly not been litigated in the public analysis of its S-1 are potentially more threatening to the company than these: Snap’s incredibly high cost of revenue per user and the fact that Instagram can so quickly and easily steal engagement from the company via fast-follow feature duplication.
First, Snap’s high cost of revenue per user. In Snap’s S-1, it was revealed that Snap has a four-year, $2BN contract with Google to use its cloud services for hosting and serving content:
“On January 30, 2017, we entered into the Google Cloud Platform License Agreement. Under the agreement, we were granted a license to access and use certain cloud services. The agreement has an initial term of five years and we are required to purchase at least $400.0 million of cloud services in each year of the agreement, though for each of the first four years, up to 15% of this amount may be moved to a subsequent year. If we fail to meet the minimum purchase commitment during any year, we are required to pay the difference.”
Snap is the biggest customer of Google Cloud; it has avoided building its own infrastructure to host and service video, which it also explains its the S-1:
“We currently have a capital-light business model because we work with third-party infrastructure partners, primarily Google Cloud, to run and scale our services rather than building our own infrastructure, which would require significant up-front capital and resources.”
These hosting costs, in addition to revenue share with content partners and inventory costs for Spectacles, are captured in Snap’s Cost of Revenue expense line item, which increased by 148% from Q4 2015 to Q4 2016:
This, of course, compares unfavorably to Snapchat’s average DAU growth over the same period of 48%. Cost of Revenue has been growing significantly each quarter for the past two years, and the Cost of Revenue per Average DAU in Q4 2016 was just under $1:
Of course, ARPU has been growing over the same period, and it exceeded Cost of Revenue per Average DAU for the first time in Q4 2016:
These unit servicing economics — while admittedly somewhat muddy, since average DAU over the quarter isn’t the same as quarterly active users — are not encouraging, and they only represent the cost of serving content to users, not acquiring them. Facebook, somewhat famously, operates data centers of its own design (and has done so since before its IPO), presumably to keep the cost of servicing users as low as possible. This is important, and it dovetails nicely with the next risk faced by Snap: that Instagram can quickly and easily steal engagement from the company via fast-follow feature duplication.
In August 2016, Facebook launched Instagram Stories, a feature that was copied more or less directly from Snapchat: users can share “stories” (videos) with their friends that disappear within 24 hours. The launch of this feature had been preceded by a long history of assaults on Snapchat following Facebook’s spurned $3BN acquisition attempt in 2013: Facebook’s launch of the Slingshot app, the introduction of disappearing messages in Messenger, the introduction of lenses in Facebook Live, Facebook’s launch of the Lifestage app, etc. (for a complete history, see this Guardian article). Mark Zuckerberg has very publicly proclaimed video to be the future of Facebook, and so it makes sense that the company would replicate Snapchat’s video-centric feature set in order to syphon Snapchat’s users away.
But compared to Facebook’s other attempts to seduce Snapchat’s users, something was different about Instagram Stories: it worked. As reported in this Techcrunch article, the launch of Instagram Stories coincided with a meaningful, widespread reduction in the number of views gleaned by videos by influencers and celebrities on Snapchat — in some cases, up to 40%.
Snapchat pithily addressed this risk directly in its S-1:
“For example, Instagram, a subsidiary of Facebook, recently introduced a “stories” feature that largely mimics our Stories feature and may be directly competitive. We may also lose users to small companies that offer products and services that compete with specific Snapchat features because of the low cost for our users to switch to a different product or service.”
But the real peril of Instagram Stories is more profound than mere feature cloning and represents, to my mind, the principle risk that Snap must overcome as a publicly traded company: that Facebook (through Instagram, which is the real Snapchat competitor) can quickly clone Snapchat’s features and then use its world class ad serving and data processing infrastructure to both earn more money per user and spend less money per user than Snapchat.
That is Snap’s existential threat: that an app that is growing faster than Snapchat can turn around on its features with more effective monetization and less cumbersome infrastructure costs. Facebook’s ad targeting machinery on mobile stands above all others: it is the gold standard of mobile advertising, with targeting capabilities that are the envy of every other ad network (with the exception, perhaps, of Google). Anecdotally, Snap’s advertiser tools are lacking, and the company is hurriedly trying to catch up with Facebook’s offering.
But Facebook has made massive inroads in bringing analytics to brand advertising placements, and the company has more resources to throw at improving its advertising tools than Snap does. And, of course, Facebook spends less per user, as a percentage of per-user revenue, over the course of serving those ads: as per Facebook’s most recent Q4 earnings report — which revealed phenomenal quarterly results — global ARPU is $4.83 against an estimated $2-3 per user in infrastructure costs (note that comparing Snapchat to Facebook here is difficult since Facebook’s content serving costs are encountered in proprietary infrastructure and are thus operating expenditure, but the margins nonetheless support the point).
None of this is to say that Snap can’t overcome these risks, but it submits itself to the consideration of the public markets in a disadvantaged position vis-a-vis a direct competitor (Facebook, with Instagram) that can quickly duplicate its features and operate them at better margins.
Recent