(I’m copying this from a long Facebook comment because I think it is worth a blog post, and I don’t really want to make a habit of giving FB only my valuable thoughts. Because you know how valuable my thoughts are.)
A friend of mine pointed to this article on the failure of the photo startup Everpix. He said he thought the failure could have been at least partly addressed by reducing their use of Amazon’s AWS. I’m not sure this is the case. From what I can see, AWS is priced very, very reasonably. For example, when I priced out Linode vs AWS, AWS was a lot cheaper and more flexible. And could replace my current colocated server at a fraction of the cost using Linode. I’m working on moving my colocated server to AWS..
Further, bandwidth for this photo hosting on the scale they would need isn’t cheap. Using AWS they can get smaller amounts of bandwidth at a very cheap price and scale up to larger amounts quickly and easily.
Finally, AWS lets them avoid the overhead of all the negotiating that they would have to do to replace each of those services.
Instead, I would like to see a desktop version of the software that would let users use their algorithm to find those important pictures faster. And then, they could have a standalone Everpix appliance at home that would take eliminate all the bandwidth and storage costs that Everpix was incurring to run the service.
I would say what killed them was the insistence on running a web-based service without trying to use any network effects (at least, I haven’t seen anyone do anything like “Here’s a list of pictures I took on this day during the last 2-3 years, according to Everpix”).
They were in the web mindset and incurred the web’s costs for hosting and bandwidth without taking advantage of any network effects. It would have worked just as well as a desktop-only application and ended up being cheaper for them.