Every organization today uses analytics in some form or another. It could be targeted towards quality and performance metrics for their processes or products, or it may have been motivated through a need from the market to provide metrics. Moving towards advanced analytics requires a significant investment and many times, this investment is difficult to justify unless there is a clear ROI derived through valuable insights. In order to derive valuable insights, there needs to be a base line definition of raw data, which is often difficult and the most expensive stage in the lifecycle of advanced analytics.
In order to justify the ROI for advanced analytics, the focus has to be on the ultimate goal. How will the business become more competitive? How will the business gain more customers through valuable insights? How will the O&M costs be controlled or reduced through process improvements? The answers to these questions will lead you on the path of a successful ROI to justify the investment into advanced analytics.
In order to proceed to the next step, vertiv recommends a few easy steps to follow:
Get everyone on-board. The first step in this process is to get everyone in the company on-board, so that everyone is aware of what you are trying to do, and can pitch in with their valuable insight and also help in identifying raw streams of data. You could do this, by making a companywide announcement, or just by sending the information to the leads in every department or vertical within the company who will in turn communicate with their teams.
Define your end goals. It is very important that you define the end state of what you want to achieve through advanced analytics and then work your way backwards to define the process or the path to get there. In order to define the end goal, you will need to identify stakeholders in every organization or department within the company and get their feedback on the issues faced and defining their dream vision.
Identify the raw data streams. Once you know your end state, you can start working backwards to identify the data sources, which could be raw data streams and any database of record within the company's network. Bringing your data into a canonical model is old-fashioned but still works. You may want to work with data scientists or ontologists to define the canonical model for your data, to make it easier for you to map raw data streams.
Scale. Focus on scalability of your data analytics, so that you can build it for the future. The infrastructure and the ecosystem should be able to withstand at least 30x of the current load as you move into the future growth of the company. Also identify the resources that you will need to scale out in the future, who can focus only on generating more analytics and integrating more sources of data.