Implementing Analytics: The Five Pillars
Big Data means that you need Analytics... and you really, really need Analytics, not just because of Big Data.
You need Analytics to:
- know what's working, what's going on with your business
- understand customers' wants, needs, experiences, and behaviors
- meet and beat the competition
Building an Analytics system sheds light on a few blazing dichotomies with your organization's capabilities.
You are a confident, skilled manager but you're over your head, overwhelmed, with Analytics. There's plenty of great analytics software and services available but most will fail to meet the results promised. You're want to do it yourself, keep control inside, but your staff lacks the technical skills.
At the start, you must have a clear purpose, set goals, have focus and organize your plan with your objectives and corresponding metrics in mind.
Analytics is complex. It's a system and a process, not a product that can be bought like a piece of software.
You must understand the major elements of an Analytics system. There are five. These five pillars of an Analytics system cover everything... and you need everything for it to work. They are:
1. Data - getting all the right data, collection, handling, cleansing, from multiple internal and external sources and silos
2. Cloud - platforms, services, applications, to get it up, running, performing and available, quickly and cost effectively
3. Processing - natural language processing of the vast unstructured text, aggregating all the different data sources and types into a common format for hand-off to Analytics
4. Analytics - purpose driven analytics, many types, performing many data interpretations, yielding actionable insight
5. Delivery - visualizing, integrating and getting the actionable insight to decision makers
Each of the five pillars is its own project, requiring different skills, technologies, services, to be performed to develop a cohesive unit that works for your purpose. Some or all can be outsourced.
All require a lot of commitment, cooperation and collaboration from all stakeholders, especially top management, to make this work. Everything needs to be measured.
In short the Analytics process needs to be managed. What does a manager do? Manage.