Writing

Software, technology, sysadmin war stories, and more. Feed
Tuesday, January 15, 2013

How did this thing get $1.5 million in funding?

I am amazed by the things which get funded as startups. I'm not talking about cat pictures here. I'm talking about stuff which might just be the simplest kind of plumbing with a web site parked on top.

One of these popped up on the "new links" part of Hacker News today. I tend to poke around in there looking for interesting things, or (more often) to flag the ridiculous spammy stuff broken people tend to submit. The title of this startup's post looked interesting so I clicked through to take a look. What I saw was nothing amazing.

So as not to identify this place, I'll use an analogy. Imagine you have a weather station, and you can look at it when you're at home. Other people also have weather stations, and they can look at theirs when they are at home. However, if you want to see your station when you're out in the world, you're out of luck because you're probably on a consumer-grade cable modem or DSL where servers are verboten. Besides, you're probably running behind one of those little plastic firewall/router boxes which makes it difficult to connect inward from the outside. Then you add in the whole dynamic IP thing, and, well, it's complicated.

This place effectively says "buy whatever weather station you want and point it at us, then you can come look at your data on our site". You can also grant access to the data to other people, like sharing the most current data or just a particular snapshot. Maybe you live in San Jose and you had a record low temperature earlier this week and you want to show off the readings. Whatever.

They also keep the readings from your weather station for up to a month. There's also a cap on how many readings they will save per hour, so you can't flood them with too much data. After a month, your old readings apparently rotate out and disappear forever.

Just thinking about this, if I were to create such a thing, I'd have a collector frontend to receive data from the different kinds of weather stations (since many types are supported, not just one) and make sure it's from an authorized user. It speaks whatever language the stations speak. For example, if they FTP in their data, then it speaks enough of that protocol to make things work. If a submission checks out, then it throws a copy of it at the backend, assuming the rate-limiting hasn't kicked in yet.

The backend is just something which sits atop some kind of storage system. Maybe it's MySQL or Postgres or something else entirely. It might even just be a bunch of files on the drive somewhere if I wanted to get really down and dirty with my implementation. It stores snapshots associated with a given client and station, and can retrieve them given similar parameters, plus either a certain date/time or perhaps "the latest one".

Then there's the visualization frontend. This is just something which does the usual "authenticate a user in a browser" thing with logins, passwords and cookies, and then it lets them choose what they want to see. It translates this into a request to the backend, and then it takes the data it gets back and renders it nicely. Add some magic so you can have a hard-to-guess URL scheme for sharing certain snapshots with people who don't have accounts and you're in business.

They want $10/month for this, plus another $5/month for every extra weather station you add on. Maybe you have a vacation house or two, and they have different weather than your normal house. Whatever. You know what really blows my mind?

This basic concept got funded for $1.5 million.

That's the amazing part to me. That kind of money would buy over 1000 years of dedicated hosting for a single whitebox (assume $120/month). Alternatively, it would buy over 50 years of dedicated hosting for 20 different whiteboxes.

I say whitebox because really, that's all you need to get something like this started. You could do all of it on a single machine for a very long time. It would have to attract a whole lot of customers and activity in order to outgrow that first box. Now, remember that every customer is paying $10/month. That means 12 customers pay for the hosting of that single whitebox. As long as they don't generate more load than it can handle or consume more than your monthly allocation of bandwidth, you win!

Obviously, you'd want more than a single box for redundancy and all of this, but if all you're trying to do is establish whether the idea will work or not, you can probably go with just the one for longer than you might realize. I saw plenty of this in my web hosting support days.

Sure, the people involved need to make money too, but really? How many people does it take to build something like this? I'm thinking the technical side of this has a very simple answer: ONE. The harder part is finding someone who is actually a big enough weather nut to have one of these things in the first place and also pay $10/month just to have someone else wrangle the data. Without sales, the technology is nothing.

So, world, consider this my announcement: if you give me $1.5 million, I will gladly build you one of those things. You want a data collector, user authentication system, storage server, and a web-based "weather" visualization frontend with "guest mode" conditional access URLs? Sure, why not. Those guys did it, so why not me?

What's the catch?