July 23, 2012

Why I think real-time analytics is a nonsense

Honestly, I don't understand what real-time analytics is. Vast majority of articles about it say that real-time analytics is basically the same as usual BI with its reports and dashboards, simply updated in real-time and that's why it's so cool. Is it?

I do understand what real-time data visualization is. I just don't see anything related to analysis or business intelligence here. Operational dashboards reflecting actual values every second or so surely are a useful tool. For operational people that act by prescribed algorithm -- sort of "on red light stop, on green light go". But how can analysis be real-time?

I used to think that analysis is an expert interpretation of data which ties numbers with influencing factors and reveals root-cause dependencies. Goal of analysis is to improve knowledge and understanding of business processes and create competitive advantage. Therefore analysis is a thoughtful process performed by human and it can barely fit into fixed time frame of several seconds (or even minutes). I believe that analysis is ad hoc by nature and presumes some data exploration, proofs for hypotheses and answers for questions. But in real-time data visualization there is no place for ad hoc analysis. Logic of the real-time is different -- either stop or go. Delay in action may cause damage, so no time for thinking -- act by prescribed instructions.

I remember the course of real-time operational systems in my university. Real-time application must produce a result in defined time frame. Result delivered one second later equals to failure and malfunction. That's the main rule of real-time systems. How can human analysis fit here?

So I think it's a nonsense.

Let's say real-time data visualization instead -- that's more correct. And let's not confuse it with Business Intelligence. This is a different story.

PS. For those who think name doesn't matter -- I'm sure you don't want to get apples when asked for oranges. Names exist for a reason.