We measured the demand from other companies and individuals that wanted to integrate with our trading platform or to create and offer a new product to the market and found it to be significant. The rough effort (finger in the air) to implement an open to anyone API service was estimated and though it would be a fair amount of effort it would be less then having to build a custom integration every time. Plus, we could use it ourselves to rapidly prototype and build new services and products. 

Building an API is not a sexy project (to most people). But we were driven to do it and to do it well.

One thing we had learned in our research is that a developer will try out an API, possibly late at night, possibly after having had a drink or two, and if they can’t get something working in under 10 minutes then they might give up. So we knew we needed to build in some instant gratification in the form of sample code and easy wins.