One could argue that the costs could have been lower by choosing some lesser or free solutions. Also it was not sure what required the $100 monthly cost to Amazon.

Still, nice to see an article from an Estonian startup.

Loved the tidbit about setting up company paperwork purely digitally.

Estonia seems to be at the forefront of e-government (you started to vote online I think 8 years ago?) and e-documents.

The cost of starting a software company and the cost of starting a "successful" software company could vary greatly. I am not having a dig at these guys, I have no idea their level of success.

The main thing that struck me was $20 on advertising. One of the biggest problems startups have is traction, some money to spend on advertising goes a long way to help with this. I am sure you can create a startup very cheap these days, but how much do your chances of success increase by having an extra $10,000 or $20,000, that would be much more interesting.

I suspect there are many startups that ran out of capital and couldn't get off the ground because of it, even though their idea and execution was good.

I personally think you need to make sure you have enough capital to get started if you are going to have a real shot at success.

Hosting in AWS are 60% of the current costs, I guess its fair to say that the cloud promise on reducing the costs is not real.

"Reducing costs" is not the same as being the bulk of your costs. In any case, AWS is still cheaper than doing it yourself.

Cloud hosting can be an amazing cost reduction, but it really depends upon a lot of factors.

When you're first starting out I believe most businesses would be much better off hosting on a small VPS at a place like 6sync or Linode. Once you reach a certain scale, the low-cost VPS model starts to fall apart and cloud hosting on places like Amazon tends to be a lot cheaper than doing the old school dedicated server/data center model.

In any case, though the point of this article is how cheap it can be to start a business these days, I think they may have preoptimized and spent too much on some things --

I would have gone with a cheap VPS solution to start while keeping a move to AWS in mind for software selection. I also probably would have gone with something like bugzilla or another totally free self-hosted issue tracker.

This is not to say AWS isn't a great value at a certain level or that hosted issue tracking systems aren't worth it, but rather that those things are probably best worried about later on once the business has proved to be at least somewhat viable.

Nice article!