"Data analysis Don'ts: Common pitfalls to avoid for Success."

I believe, in today’s era of Augmented analytics companies are inundated with predictive and prescriptive data models. In addition, usage of sophisticated continuous intelligence tools (CI) is quite common. To use these data pointer effectively there any couple of cautions beyond technology which we shall be vigilant.

1. Avoid ‘Confirmatory bias’s: Many times, it happens that we are so inclined towards a particular hypothesis that we scout for a selective data point that supports our hypothesis. Data will always tell a story which we want to listen. Therefore, do not analyze any KPIs in isolation. Use a stipulated formats to avoid these biases.

2. ‘Presenter x Reviewer contradictions’: Data is a beautiful piece of art and same data formats can have multiple inferences based on the role of data user. While we are presenting the data, most of the time we highlight only those points that underlines extraordinary performance. While, when we are reviewing the same data points, we exhibit a critical outlook. To avoid these contradictions, we must also give weightage to analysis provided by data analysis team.

3. Follow the sequence

·?????? Observation i.e., Data collection from multiple data sources

·?????? Analysis i.e., Data articulation and Statistical modeling and

·?????? Insights i.e., Synthesis of analysis, connecting dots and Data presentation to derive inference to formulate a successful strategy.

It is important to follow a sequence while formulating strategies and do not practice each block interchangeably. Many times, due to urgency at work, we directly move to either analysis or Insights which will result in missing critical success factors.

4. Analyze data even when performance is extraordinary.

It is said, "Repair the roof while the sun is shining."

Most of the time, we deep-dive into data when performance is not akin to the planned objective. However, it is more important to conduct a detailed study to understand the critical success factors. So that the success is sustainable, and the learnings can be replicated.

and last but not the least

5. Data is not just about numbers

It is said, "Not everything that can be counted counts, and not everything that counts can be counted.”

Similarly, each KPI might not be available in number format. There would be qualitative parameters such as expressions, opinions, feedback that can help us to understand what users need, why problems occurred, and how to solve them.