Kaizen, the manufacturing movement for continuous improvement, is being adapted into many different disciplines, including analytics. The core tenements of Kaizen are to
- Start small.
- Remove overly complicated work.
- Perform experiments to identify and eliminate waste.
There is an emphasis on delivering value quickly rather than completeness. Testing and learning can make small improvements along the way while working toward the end goals.
This is a marked contrast to the current tendency to spend long development cycles building the “perfect” models. Today, building and deploying analytics are complex, custom projects that cross over multiple functional areas. In this new era, modern analytic teams are removing the ivory tower academic shackles from traditional analytic methodologies to eliminate unnecessary, time-consuming steps from the project cycle. This helps increase agility and responsiveness while incorporating business feedback into the process so results can be improved.
With Kaizen serving as a guiding principle, modern analytic teams build and deploy models immediately and then improve on the models in short burst cycles dovetailing the work of analysts and IT to create a frictionless environment that continuously delivers business value. On the ground, the teams often use hybrid agile or rapid application development methodologies to improve cycle time and reduction in barriers with cross-functional teams.