"The $64,000 Question" began as a radio show in 1940, and went through a series of transformations and ultimately, adoption to television until it finally left the air in 1952. Contestants were asked questions of progressive difficulty as the stakes of the game increased. I still hear people who weren't even alive at the time of the show refer to the "$64,000 question" when they talk about the art of asking the right question and getting the answer that they're looking for. I dare say getting both questions and answers right hasn't gotten any easier with big data.
That is the question
The point was driven home this week in a conversation I had with Srikanth Velamakanni, CEO and founder of Fractal Analytics, a big data analytics provider. "What we see with companies and how they are using their big data is actually a bigger issue than the answers that they are deriving from their big data," said Velamakanni. "We believe that the big issue is whether they are asking the right questions in the first place."
It's a challenge for organizations which now find they are swimming in data, yet struggling to identify the really relevant questions they should be asking.
Some organizations try to harness their big data by structuring analytics questions around specific business cases that they want to solve or better understand. The thinking behind this is that if you have specific objectives and a tightly constricted focus, you won't get "off course" in your big data probes and questions asking, and you will likely arrive at results faster.
Yet, most of us only have to go through a current report catalogue in the average company to see that there already is information on company business line and product profitability, quarterly and annual financial results that can be compared with financial performance one year ago, statements of inventory surpluses and shortages, and even reports that show how many new accounts the company has gained (and how many it has lost). These reports are the outcomes of historical business cases that companies have already identified in areas like finance, operations and sales - so if a big data analytics project gets focused too tightly around one of these existing business cases (or something similar) the chances are high that the answer to a big data query (or at last part of it) already exists in legacy reports somewhere.
Art of the question
Velamakanni talks about this when he discusses the art of asking the right questions and getting the most out of big data.
"We have scientists in our data lab who tell us something we were not expecting to get," he said. "It comes down to the ability to explore data. If you don't do this data exploration well enough, you might find that you're too narrowly focused on solving a particular business problem with your queries. On the other hand, when you look at the data itself without initially limiting it with a focus that could be too narrow, you might come away with something different."
The exploratory approach works well in academia, where the emphasis is on research and there is also an understanding that not every data exploration will be successful, However, in enterprises where bottom line results are measured and bottom line answers to questions are expected, it isn't always easy to adopt a more open and exploratory approach to big data queries - especially if they don't end up yielding results.
Of course, if you don't dare to experiment with big data, you're not likely to get the ground-breaking answers you want to get, either.
This is the crossroads that many organizations find themselves at now in their big data analytics. How they achieve a proper balance between business case-driven results and pure data experimentation that yields innovative, unanticipated and actionable information - might well be the next "$64,000 Question."
0 comments:
Post a Comment