Whenever I begin an analysis problem on a new consulting engagement, I start with the same questions to focus my understanding of the problem, the dataset and the culture of the organization. Having a standardized flow helps me quickly sink into the analysis and end up with a solution that adds real value, rather than just creating some pretty dashboards (or worse, redoing cross tabs from Excel within Tableau). This article is the first in a series of posts on the Archetypal Analytic Cycle ™ and its applicability to a wide variety of applied analytics problems, especially visual analytics problems. So, without further ado, here are the questions I begin with, in the typical order I try to answer them:
- What is the problem to be solved?
- How does the client (individual) get bonused or remunerated (i.e. what specific levers are they being measured on)?
- What are the strategic goals of the organization, if public, from their annual report, or, if private, from interviewing executives?
- What is the source of the data?
- How is the data shaped?
- Can the problem be solved given the data source and shape?
- What story is the client (individual) trying to or has been telling?
- What story is the data trying to tell?
- Are #7 and #8 in alignment?
- What is the analytics capability and maturity level at the client and business level?
- What are the client’s analytic biases?
- What are the organization’s analytic biases?
- What are the domain analytic biases?
- What are my own analytic biases that I bring?
- Is each component of the overall analytics deliverable actionable?
Although I am biased to using Tableau as a visualization solution (I do work for Tableau after all), I answer all of the above questions before even beginning data exploration in any tool. My preferred method is to whiteboard the questions and answers before even opening a laptop or using any tools. Then, when I begin the actual data exploration phase, I have some confidence that the project is not an Alice in Wonderland down the rabbit hole adventure.
In the rest of this series, I will explore each question in depth, diving into sub-questions that inevitably arise as the main questions are asked and answered. Also sprinkled in will be case studies that have been modified to protect the data and organizations involved.
‘Til next time!