Analytics Life Cycle(ALC)
For those not yet in the thick of analytics action, but aspiring to be there, the implementation of analytics by organizations to support decision-making may sound like a complicated, an often daunting activity. Is it data base management systems (DBMS), is it python or Hadoop, is it SAS or R, is it data visualization and Tableau, is it data extraction and modeling, is it business intelligence and data science, is it a lot of statistics? Or it is ALL OF THESE? And more!
For the uninitiated, stepping into the world of analytics can be like the classic parable of the blind men and an elephant, with each of them arriving at a different definition of what they were touching because of the different places where they were standing.
Making sense of the ‘elephant’ that analytics could be, therefore requires a unified approach. The Analytics Life Cycle (ALC), presented below, provides a framework that is easy to understand even for someone with no prior background in analytics. The first step in the ALC is problem formulation, whereby the business users (manager) specifies the issue/opportunity they are facing, e.g. customer attrition, falling margins, or declining sales. Next, the business analyst will gather and organize data that is required to address the business need (step 2). After initial data exploration and transformation using statistical techniques (steps 3, 4), analytics tools like SAS or R will be used to build and validate the model (steps 5, 6). Finally, the model is deployed and its impact monitored in a ‘live’ business environment e.g. in an online shopping environment to drive average transaction value (ticket size).
From the ALC, it must also be evident to us that successful analytics implementation requires a diverse range of skills to come together, combining statistics, data science, programming and database management, and business/domain knowledge.
This has given rise to the concept of “PURPLE” skills…about this, we will talk in another blog. Keep reading!