I left the TDWI conference in Las Vegas a bit over a week ago with my head afloat in buzzwords. It seems that every quarter there are new terms and that I was not the only person at the conference struggling with the ambiguity and apparent overlap among terms. Maybe its time that we define our terms and differentiate between similar sounding terms.
I've listed below some of the language that stands out as needing clarification. Along with each I have included a tentative description as a starting place for some discussion and clarification. Please weigh in with your thoughts about the meanings of these terms and add to the list any that you believe need similar attention.
I've listed below some of the language that stands out as needing clarification. Along with each I have included a tentative description as a starting place for some discussion and clarification. Please weigh in with your thoughts about the meanings of these terms and add to the list any that you believe need similar attention.
- Analytics - Many people use this term to describe analytic tools and the things that they produce (dashboards and scorecards). I think analytics includes the entire process of mathematical and statistical analysis to gain and use insight.
- Embedded Analytics - My interpretation of this term is that it describes analytic capabilities that are tightly coupled with a hardware platform ... the basis of analytic appliances.
- Operational BI - Business Intelligence feedback in the form of reports, dashboards, and scorecards into the operational processes and applications of the business enabling a strong connection between strategy and operations.
- Operational Analytics - Integration of analytics (see definition above) into operational systems. Operational systems in this context include both planning systems (supply chain, demand forecasting, sales planning, etc.) and transaction systems (order processing, supply chain execution, etc.).
Posted March 10, 2009 5:41 PM
Permalink | 1 Comment |
Twitter 
LinkedIn 
YouTube
I'd agree with the definitions and would add that operational analytics takes full advantage of all the data, no matter how large, to increase the precision of the analysis. Think about evaluating hundreds of attributes rather than a dozen or less. Addtionally, I see operational analytics as in-line or streaming with the operational process to very quickly increase the intelligence of the operational system/process. Both of these cases are much more reflective of the real world. By being able to evaluate the tails or outliers you move from a reactive position to a proactive position of identifying emerging trends. The embedding of analytics speeds the process of performing operational analytics and allows the processing of complex analytics on mainstream hardware. In the past, companies didn't pursue the really hard complex problems on large data because the computing power it required was prohibitively expensive. With embedded analytics, companies have the opportunity to pursue a whole new class of opportunities to unleash pent up business potential.