The application of data science is generally the investigation and testing of a hypothesis to uncover an insight or attempt to solve an existing problem. The reality with any such testing is that the identification of value is not guaranteed, with the risk that if left open-ended can result in resource consuming, costly, and never ending analysis.
To reduce this risk, we approach this through the use of time-boxed ‘proof of value’ (POV) scenarios. With clearly defined timeframes and success criteria, we can prove potential value before moving forward into implementation of repeatable, and actionable insights.
The goal is to leverage existing investments for implementation – such as HANA or Predictive Analytics, rather than ongoing execution by individual data scientists