The last few years have seen algorithms taken on an increasingly expansive range of predictive endeavors, ranging from predicting the price of oil and the best places to position police officers all the way to the chances of success for blockbuster movies.

A recent venture has attempted to use big data to perform similar predictive miracles, but this time in forecasting whether a startup would succeed or flunk.

The project, called Test4startup, uses an algorithm that was developed in conjunction with a university professor to supposedly test the prospects of a startup idea.

How it works

Users describe their business in as much detail as possible, and the algorithm then creates a pool of keywords that reflect the genesis of the idea.  This is used to define the market and establish the potential for that sector.

The system then cross references this with the trends it has identified for that sector by mining the news and various other sources to build up a picture of the health of a market, and indeed the kind of rivals you might face.

It then compares your data from these two steps to compile a list of recommendations for you and your startup.  It might tell you, for instance, if you’re pricing your product too highly and that you need to reduce it by $x to properly compete with your competitors.

Keys to success

It’s an interesting approach, and I’ve covered before that there are tools emerging that attempt to automate the kind of horizon scanning that the tool aims to do when analyzing the state of the market.

It does however assume that the idea and the market conditions are the key determinant of potential success.  The big elephant in the room is the lack of any kind of analysis of the startup team.

The importance of this team is something I’ve covered a few times, and it’s something that is always going to be difficult for an algorithm to assess, despite attempts being made to do just that in recruitment scenarios.

Nonetheless, it’s an interesting project and one that will be worth keeping an eye on.  Whilst I can’t see it taking the place of experienced investors just yet, it’s quite possible that algorithms will play a growing role in investment procedures.