Predicting the Outcome of Startups: Less Failure, More Success

On an average 9 out of 10 startups fail(industry standard). Several reasons are responsible for the failure of a startup including bad management, lack of funds, etc. This work aims to create a predictive model for startups based on many key things involved at various stages in the life of a startup. It is highly desirable to increase the success rate of startups and not much work have been done to address the same. We propose a method to predict the outcome of a startups based on many key factors like seed funding amount, seed funding time, Series A funding, factors contributing to the success and failure of the company at every milestone. We can have created several models based on the data that we have carefully put together from various sources like Crunchbase, Tech Crunch, etc. Several data mining classification techniques were used on the preprocessed data along with various data mining optimizations and validations. We provide our analysis using techniques such as Random Forest, ADTrees, Bayesian Networks, and so on. We evaluate the correctness of our models based on factors like area under the ROC curve, precision and recall. We show that a startup can use our models to decide which factors they need to focus more on, in order to hit the success mark.