Cost of Bad Data

According to a Gartner study, bad data cost an enterprise between $9.7 to $14.2 million every year. This translates to a potential reduction in revenue by 30% and an overall estimated cost of more than $3 trillion dollars to the US economy. For a startup, where every decision can be very impactful and any negative impact to revenue due to bad data, can make or break a startup. Whether you are bootstrapping or have investment, bad data can increase your burn rate by making it difficult to make the best data driven decision.

Focusing on anything about data is something that is close to my heart. I am a data nerd! My background is in statistics and for about a decade, I have been working in machine learning and AI helping Fortune 500 companies and large government agencies deploy machine learning at scale.

I want to share a story of a time I had to interact with some bad data. I was working on a machine learning project for a big institution where we were working on a fraud detection use case. While working on this project, I came to realize a lot of the data was spread throughout the organization and it was difficult tracking all of it down. There was a lot of inconsistency in the data and it was business critical data. As a result, it required many hours of manual work to be able to organize and clean the data. It took months to get everything to the point of building a machine learning…