I see this as a talent/technology grab instead of a revenue play. Obviously, Skype will be able to integrate the technology and find a way to monetize it with their user base. There have been multiple companies purchased before they even launch a product. What's the difference?
The fact it has users actually makes it a negative-revenue company, since every message the users send costs GroupMe money. At 100M msg/mo, that's probably a bit more expensive than keeping some servers running.
That's less than 40 messages/second. Even if every message goes out to an average of 20 people, that's less than 800 messages/second. If each message averages 1KB (and why would it be that large?), 800KB/sec, or less than 7mbps.
For high availability, I'd guesstimate maybe 6 physical servers in geographically diverse locations.
If they're spending more than a couple thousand a month on raw infrastructure, I'd be utterly shocked.
Edit: By the way, this kind of thing is nothing new. IRC networks have handled traffic with similar properties for 20 years.
GroupMe was sending SMS text messages, and while they're nominally just a tiny amount of data, the cellcos don't bill the same way as your local upstream internet provider.
Correct. So... use that 800/sms/second. Even assuming they were getting SMS at a bit over .1 cents per SMS (highly unlikely - probably much higher than that), that's about 80 cents per second - round it up to $1/second that would be $3600/hour.
Sheesh - is my math right on that, or did I move a decimal place?
I'm guessing they weren't burning through thousands per hour every hour on SMS fees, but even hundreds would put a serious dent in a startup's pocket.
