Dec 11, 2013
Developing Big Data Analytics Strategy
The original steps
The author of Developing a Big Data Strategy article identifies 6 steps for developing big data strategy. He focuses on the big data but these steps are necessary for any focused analytical effort in any organization. The steps with their short description are:
Step 1: Comprehensive Data Inventory.
This step involves making detailed inventory and understanding data available on hand.
Step 2: Strategic Business Requirements Gathering.
At this step an intimate knowledge is gathered for the business process, the ways data is used and the questions being asked across organization.
Step 3: Existing IT Infrastructure and Software Inventory.
Evaluation if existing IT capability can meet the new requirements coming developing the analytical capabilities.
Step 4: Build Business Case
The business case involves comparing of costs and benefits of the (big) data analytics.
Step 5: Develop Big Data Capability - Infrastructure, Metadata, Software, BI (Reporting)
This is the stage for building the infrastructure and other tasks identified on previous steps as well as a prototype of the whole process.
Step 6: Build Analytics Capability.
This step is about filling in the analytical ranks and start producing analytical results identified on earlier stages.
Please read in details here. The steps answer to the questions:
- What data do we have on hand?
- What is important in our operations and strategy?
- Can current IT infrastructure meet the new requirements?
- Would advanced analytics bring value to the organization?
The author nails some important points. However, I find the steps need to be reordered and enriched with some activities.
I would put step 2 on the first step of this process - the all process starts with the need for developing new analytical capabilities in the organization. Most of analytical needs are covered in one way or another and Step one would also include evaluating the cost/benefits of the current way it is done.I would also add that aligning on strategic goals, plans for new products, services, markets, etc should be organic part of this step.
I would include a research on new data sources in current Step 1 as some data generated in the business processes is simply not collected or properly stored. Acquisition of external data should also be considered as it could bring light on customer behavior, demographics, etc. Data could also be generated through surveys. Requirements for data come with the analytical goals and are connected with the findings on previous step.
I would include the development of model how would all work on step 4. I would make this point more explicit as as analytical processes and results should organic part of the business processes in the organization.
I would leave the rest of the steps on the list as they are with two important notes.
First is that all steps should be interactive and going back and forth is necessary.
Second point is that once built the analytical framework is not carved in stone - the organization is a living structure and its analytical needs change with the time.
There are many ways to approach development of a (big) data analytics strategy. I believe the the key for a successful strategy is keeping cool head, ask the right questions, focus on the real needs and have future in mind.