Today, with the focus on big data and technologies like R and Hadoop gaining prevalence, it seems that many organizations think that all they need in order to be successful with advanced analytics is to hire a few “data scientists.” The intention of the “Myths and Realities of Successful Analytics” series is not to rant on this now-overused word for a skilled profession (utilizing techniques that have been around more than 200 years). It is just that an organization needs more than a few people that can roll code in order to be successful implementing advanced analytics solutions in their organization.
Businesses that have been successful with advanced analytics need:
- A well-defined business problem.
- The identification of appropriate data.
- Manipulation of the data.
- Preparation of the data for modeling.
- Analytic skills and technology to develop the models.
- Validation of the results.
- Deployment and integration into operational systems.
- Monitoring and updating of the modeling efforts.
Over the past 15 years, I have worked with both commercial and public-sector organizations helping them identify, design and implement successful advanced analytic solutions. In this series, I hope to set the record (a little) straight(er) on what it takes for an organization to be successful implementing analytics. I by no means propose replacing tried and true analytic methodologies like SEMMA, CRISP-DM, the KDD Process, or DMAIC, but rather this series serves as a framework you can use to develop an advanced analytical strategy.
Follow the Business Analytics Knowledge Exchange for weekly installments of this series. Next week, I’ll address the skill sets and people you need to build your analytics dream team. Until then, check out this nice article talking about how big data is not enough to be effective with analytics.