Every time someone tells me they are hiring a data science consultant a tiny voice in my head screams — wait what does THAT MEAN. A data scientist is supposed to interpret your companies’ data to explain what that data is saying about your business based on how your business operates. The first step in this process is to make sure the person tasked with this job understands your business completely and what your data means. So hiring a consultant on a 6 month project makes no sense because the consultant will be learning what the organization does for 6 months. Data science is the interpretation of data in its relevant context. For example a doctor is technically a data scientist for the body. They use data like tests for certain factors in combination with their knowledge of the human body to understand what that could mean. They go to a lot of school to understand the human body. Hiring someone to just read the tests without knowing the context misses the point. A data scientist who does not understand your business is a Physicist who does not understand physics. I accidentally started doing “data science” in 2012 because our company was trying to figure out how we could make more money. Back in the good old days we called this data analytics. Most companies want to understand what their data is telling them and to do that they now hire data science consultants. This approach fails its most fundamental test- they probably don’t understand how you operate your business- so they dont know what any of your data actually means. The best way to interpret your data will always be to find someone who understands how your business operates, then triangulate the data and map it to your processes. You don’t need to hire more nuanced titled people to basically get the advice that “you’re not collecting the right data”. Because if you are now in the rabbit hole of “collect more data that looks like x” then your data strategy might look cool, but won’t be reading your business properly. You can’t infer insights when you don’t know your subject matter. You don’t need a fancy degree to do data science; you need to know how a business operates and have a functional understanding of how to put information together. The most important part of data science is understanding how a business is run. Then you can learn how to save your company $5 Million in a year by NOT doing something.
0 Comments
|
AuthorArchivesCategories |