According to Chris Dixon, a full stack startup‘s goal is to “build a complete, end-to-end product or service that bypasses existing companies”. Tesla is a good example. The company could have chosen to become a car-battery manufacturer. Instead of that it went for the fences and achieved the much harder goal of becoming a car manufacturer. In the process, Tesla can deliver a great customer experience, doesn’t have to negotiate with older companies and control its margins. All of this because it has chosen to go full stack.
What is a full stack graph startup then? A full stack startup derives an unfair advantage from its graph technology. This unfair advantage can materialize in industries :
- that are data oriented ;
- where data is highly complex and connected (ie hard to model with traditional tabular oriented databases) ;
- and where great value can be derived from the connections in the data ;
As a result full stack graph companies tend to focus on a few key sectors like logistics, social networks or marketing. The most successful full stack graph startup probably is Facebook. As early as 2007, Mark Zuckerberg presented Facebook as a “Social Graph” that links people, tastes, places, objects, etc. Facebook relies on that graph to monetize its marketing platform. The data challenges that come with building a social graph are enormous : the social graph is a large dataset (as in very large), made of connections between entities and the value lies in analyzing these connections quickly.
To solve this problem, on its way to become the company it is today, Facebook has designed a graph-oriented data structure called Tao. Facebook has also backed Giraph, a graph computation framework, that it uses to understand a trillion-edge large graph. By solving these tough problems, Facebook is able to deliver products like “Graph Search” other social networks (I’m looking at you Google+) would struggle to come by.