This is one of the ugliest financial statements I've ever seen. The multi-billion dollar loss is due mostly to stock compensation, but still their COGS or cost of revenues sold was greater than their revenues. That's probably the first time that's been seen for a sizable software company in the history of the industry.
They said their revenue per DAU is $.90 and their server costs per DAU were $.60...so that means server costs were around $100mn while total COGS were $160 mn. So what was that other $60 mn spent on? Admins? But I thought they used the Google App Engine infrastructure to avoid any server admins and such.
Their revenues are exploding with their revenue up nearly 400% from last year, but were below estimates as was user growth which are probably causing the current stock price drop.
"server costs" (Google Cloud) is an affordable $450 million per year commitment.
They apparently have 2,360 employees doing who knows what (not growing the user base).
My favorite part:
"Spiegel laughed out loud when asked if he was worried about Facebook copying his features. “Just because Yahoo has a search box doesn’t mean they’re Google.""
We'll check back in a year to see if he feels the same way.
> They apparently have 2,360 employees doing who knows what (not growing the user base).
I don't understand what all these people do all day. Same with Twitter, they have 3860 employees according to their website. It seems like 10X as many as needed. I've worked in a few offices with less than 100 employees total, and while I have no idea what it takes to run a service with that much traffic, Snapchat runs on Google's cloud service.
> “Just because Yahoo has a search box doesn’t mean they’re Google.""
Maybe it will be a classic in the same genre as Ballmers answer to what he thought about Apple's new product the iPhone.
I wonder if they would have done better with a stack like WhatsApp - Erlang running on their own servers. Or even switching over to that in the future. I think WhatsApp at 500m users had about 35 engineers and 550 servers and so costs maybe 10% of Snap's. http://highscalability.com/blog/2014/3/31/how-whatsapp-grew-...
Bandwidth will almost certainly be their dominant economic input.
It's funny, a lot of people think you exit AWS/GCE to rack your own machines, but I was on a datacenter tour in SV a few years back and was reminded that having fine-grained control of your network is just as a big a reason to do it. Adtech/finance do this all the time to control latency, I imagine it's also a consideration for Snap.
I wonder if they'll ever build a CDN. It's probably not the right traffic pattern without a lot of re-sharing, but still, I could see it.
I read their S-1. The other problem is that ex-US, bandwidth is more expensive than in the US, even though the addressable ad markets are smaller. People are poorer and don't spend as much.
You have to admire facebook. They've been focused on low-resource Android for a long time, even requesting employees to use slower connections from time to time to make them viscerally feel the effect of low-bandwidth Internet. fb really gets that they need an Asia growth story and they've done pretty well. I simply don't see how Snap will do anything comparable given that their entire service is predicated on high-bandwidth video shot from high-end smartphones even half the US probably can't afford.
They said their revenue per DAU is $.90 and their server costs per DAU were $.60...so that means server costs were around $100mn while total COGS were $160 mn. So what was that other $60 mn spent on? Admins? But I thought they used the Google App Engine infrastructure to avoid any server admins and such.
Their revenues are exploding with their revenue up nearly 400% from last year, but were below estimates as was user growth which are probably causing the current stock price drop.