11 lessons for doing a successful data analysis

Analysis has lot to do with what you ask than what you deliver. Or in words of Stephen R. Covey: “Seek first to understand, then to be understood.”

There’s no such thing as a perfect answer. You must have heard this a bazillion time. And it is true.

Well, it is especially true in case of analyzing data.

Because there are no bad answers, only bad questions.

Over the last many years, i have failed, learnt, tried and succeeded in working on some of the best analytics projects.

Regardless of the title I held, my goal was always to enable an action. So below, i am listing a few of my learnings over the years in the ever-dynamic field of analysis.

1. It is not the results but the way analysis connects to the problem in the broader context clearly enabling the stakeholder to take a decision.
2. Understanding the statistical construct of the data, and the biases involved in the question delivers a robust analysis.
3. It is the clarity of question that drives an analysis.
4. Interpretation of data depends on how well you understand the objective of your stakeholder.
5. Simplicity in the results drives an action, produces confidence in the analysis.
6. Analysis is not complete until it supports a next step.
7. Your role is not to make data ready or produce fancy charts, you are responsible for making it usable.
8. The role of helping stakeholders to ask the right questions is very undermined.
9. Break your problem in its fundamental and basic parts, to address it clearly.
10. Use easy to understand language, refrain from using jargons that help no one.
11. Evaluate the problem from top down for understanding how it affects or gets affected by the broader organizational context.

And remember, you are not responsible for making the fanciest models and presenting cool charts. You can do that totally.

But your primary role is make the findings into an insight instead of “an interesting observation” that’s not really useful or implementable.