Being able to tell a good story is as important as good data analysis.
Sherlock had the ability to analyze and put together odd pieces of events. But his findings were set on paper in a catching way by his friend Dr. Watson. Being able to tell a good story is as important as having the right data analysis to back it up.
Traditionally, concerns were focused on how to minimize data processing time and how to build a model with the highest predictive value. Today’s concerns are towards what actions can be taken based on predictive modelling and what constituencies will support or block implementation.
“Data, hardware, and software are available in droves, but human comprehension of the possibilities they enable is much less common.” Tom Davenport. HBR.
Data analytics is without question on the rise and it was enabled by technology. Today there are thousands of businesses that collect vast amounts of data but are at a loss when trying to put this information to use.
Why are both important?
In an organization efforts are aligned through its strategic objectives and in most cases data allows to measure progress in order to reach these objectives. Within the decision making processes in organizations people with tech and non-tech skills coexist and both are equally important. “Without data you’re just another person with an opinion.” As said by Deming. But without a hypothesis/objective you’re just a person with data. Communicating proposals validated on data points and generating consensus throughout the organization drives meaningful new ideas. Enabling to leverage data in order to achieve business results and create insight.
When does analytics fail? The case of Netflix
A couple of years ago Netflix launched a $1 million prize for the team that could come up with an algorithm that improved by 10 points the current match making of recommendations. So the algorithm was developed and there was a winner but it was never implemented because Netflix changed its service from DVD-by-mail to streaming. Meaning the whole organization was changing and the algorithm developed was rendered useless in most part.