Considerations for the platform of a new startup

Strategy, Architecture & Problem-Solving

Considerations for the platform of a new startup


I’m seriously considering building my new startup on Microsoft’s Azure platform with the BizSpark programme and that’s a departure for me.

I’m going to use the term “startup” loosely but at least 1 of the 3 conformed to Steve Blank’s definition, proving the business model worked. I’ll cover three of the ideas; 2 built and 1 about to be built.

First Startup

I’ve built my first startup on Amazon’s AWS. The 3rd party documentation on how to configure a linux instance to accommodate Ruby on Rails was a key driver in that decision. Incidentally, I was also new to Rails, so I had a double learning curve. The documentation was thorough enough and up-to-date enough that I was able to follow the instructions and maintain the instance for a year.

I didn’t like the charging mechanics of AWS for startups, not boot-strapped startups. My idea at that time was large and could only start large. There was no concept of building it small and growing. It had to start large and grow from there. So the small instances offered in AWS startup package weren’t sufficient. So that meant projecting costs forward 12 months and deciding on reserved versus on-demand usage, etc.

That led me to rethink my next startup

Second Startup

My next one was much smaller. More an attempt to keep my knowledge current, in terms of startups and development. I’m not a developer by profession (or haven’t been for many years), so the same as playing guitar, if I want to be able to perform, I have to practice. I chose Heroku. Partly because I wanted to build at no cost and partly because I wanted to explore the limits of Heroku (I found them). In the first startup, I had Apache/Lucerne Solr and found that I’d need to licence this separately as a service on Heroku. There were a few options, all by 3rd party developers.

That made me consider Digital Ocean, etc. Not for a full-live release, but while the startup is proving its business model.

Towards the end of the first startup, I encountered Google Cloud Platform. That had some real advantages that I liked the look of. I really like the pricing mechanics in that the cost you pay was calculated based on your usage. The more you used it, the lower the unit price. So you could end up with a similar cost to AWS’ reserved leasing, but with the on-demand commitment. I liked that idea. Plus I began to understand that Google Cloud Platform offered similar server instances to AWS rather than just the app engine (which had the reputation of only allowed certain languages); something that I hadn’t realised before. So from my perspective, GCP became my first option for a new startup.


So what changed my opinion towards Azure? I’ve been thinking about Azure for a while. In terms of whether it would make connecting to business applications easier than with the other platforms. I’m sure the others would work, but I’ve been considering reducing friction to change. Whereas my previous startup concepts were B2C (although the first had a significant B2B element in it), this current idea would be B2B so I want it co-exist with the ecosystems of clients. And most of them will have Windows operating systems and applications. The Gartner Magic Quadrants, including that on Cloud IAAS positioned Microsoft as a Leader in the quadrants.

Add to Azure, the support from BizSpark and it begins to look an attractive proposition. At least one, I’ll be happy exploring for this next project.

Print Friendly, PDF & Email