Book: Creating a Data-Driven Organization

Author

Carl Anderson

Summary

A blueprint to create a data-driven and analytics focused organization.

Takeaways

Data-driven organizations are more successful and generate more value through better decision making.

For a company to be data-driven it must have the right culture and talent in place to use data effctively along the so called analytics value chain. Sponsorship from high profile positions in the organization is needed to implement a culture that values testing and experimentation to derive insights that can give a competitive advantage.

In a data-driven organization, relevant data of high quality feeds reports and stimulates deeper analyses that are presented to decision makers who incorporate them in their decision making to influence the direction of the company.

Quotes

“A data-driven organization will almost certainly be choosing among future options or actions using a suite of weighted variables. Resources are always finite, and there are always pros and cons for different reasonable courses of action. One should gather data for each of the set of variables that are of concern or interest and determine weights among those to generate a final leading decision."

“Data is the raw, unprocessed facts about the world. Information is captured, processed data, while knowledge is a set of mental models and beliefs about the world built from information over time."

“This pseudo-progression is often labeled as analytics maturity. If you do a Google image search for ‘analytics maturity’, you will see what I mean; that many BI vendors and practitioners present this as set of stepping stones with unidirectional arrows pointing from one level to the next. Analytics is not like that: it cuts across levels within an analysis, and different parts of the organization can be engaged in analyses of differing degrees of sophistication at any one time."

“The key here is to start with the question to be answered—be question and decision focused rather than data focused. By setting out the objective clearly and unambiguously, you stand a better chance of defining which questions need to be answered and consequently which data should be collected, which experiments should be run, and what metrics you are trying to drive."