Business intelligence (BI) introduced a new way of thinking about processing and surfacing data. Analytics seems to be simply an extension of BI, but in many contexts, it's a way of thinking as far in advance of BI as BI is in advance of old-school transaction processing.
At face value, this difference is hard to see — both approaches exploit big data, and both are used for decision support. Analytics might even be seen by a casual observer as just another name for BI, but the difference is huge.
BI is about analysis, and analytics goes beyond analysis. BI refines existing business solutions using known data relationships to select from among options, and is generally linear; analytics is multi-dimensional, and draws its power from pattern finding beyond known relationships or current expertise. In short, BI makes us better at what we already do; analytics makes us better at what we could potentially be doing.
Someone who builds analytics solutions needs to possess specific skills. I list abilities that are key to being a successful analytics solution designer.
1: See various ways the future might unfold
Though predictive analytics is specific to a sub-domain of business problems, most analytics are predictive to some degree. Operational analytics — often designed to exploit real-time data streams — are about detecting the immediate zig or zag in trending information, for instance.
The analytics solution designer is, to some degree, someone with a knack for seeing different ways the future might unfold. This translates into an understanding that success in analytics is resisting the impulse to configure a solution in such a way that it only services one general outcome.
We already have that — it's what BI does. Analytics take us beyond general outcomes into unexpected new ones. From a practical standpoint, that means configuring solutions that are overtly open-ended.
2: Know all the data sources
Moreover, getting the inputs right up front can have serendipitous results, including the discovery of unexpected trends in the data and a more extensible solution architecture.
But the nature of big data is that expansive search and computational refinement carry a cost, and that too much data works against the solution. Analytics says, philosophically, no, that's not true — that's BI thinking. The trade-off is that even though More is Better, in pattern-finding for effective predictive solutions, the computational cost of throwing everything into the pot is that it must simmer forever before it turns to soup.
An effective analytics solution designer knows the structured and unstructured data sources, and has developed the technical skills to understand which ones represent the heavier computational burdens. The knack for including fruitful data sources in a solution built to discover new outcomes is a strong intuitive grasp of the balance between practical interactions and efficient interactions in pattern-seeking.
3: Scope a business decision support solution
On the other hand, while a liberal approach to input is a plus, it's also important to keep any individual solution — especially in a real-time system — focused on solving a particular problem, or supporting a very focused decision point in a business process. Why? Because making a predictive solution too broad introduces ambiguity in the outcomes.
A car manufacturer wouldn't build a solution predicting customer response to all the vehicles it makes — there are too many variables, vehicle to vehicle, that would affect different customers in different ways. That principle applies for large and small decisions.
This translates into a reasoned and a seasoned ability to correctly scope a business decision support solution. The successful analytics expert will take any given decision support problem and tease it upward and downward until an optimal focus for the solution is identified.
Analytics solutions are becoming a specialty that will take root and grow in the coming years. Cultivating the appropriate skills for this role is a new frontier, and those who undertake it are in for an exciting decade.
- Remember FLURPS to design better big data analytic solutions
- Data scientist: Your mileage may vary
10 curses of the analytical thinker
- Birst: Business intelligence and big data insights for the everyman
- 10 things you should know about big data's secret sauce
- Job description: Big data modeler (Tech Pro Research)
Note: TechRepublic and Tech Pro Research are CBS Interactive properties.
Scott Robinson is a 20-year IT veteran with extensive experience in business intelligence and systems integration. An enterprise architect with a background in social psychology, he frequently consults and lectures on analytics, business intelligence and social informatics, primarily in the health care and HR industries.