Big Data

Big Data, Big Answers

on September 30, 2014

John Lucker, Principal and Global Advanced Analytics and Modeling Leader at Deloitte, describes breakthrough analytics that transform strategy (and assist operations) as “crunchy” questions because,  "They are capable of forming the connective tissue between tactical and senior [strategic] level objectives in an organization."  Lucker suggests that instead of asking her team about current market demographics and sales results over the past six months, the Chief Marketing Officer asks instead, "Who are our next 1,000 customers going to be, where are they going to come from, and how and why are you going to win them?" The answer lies within the analysis of crucial data. I like to call the results data analysis can provide as “Big Answers.”

In Mary Shacklett’s Big Data blog, Lucker continues by saying, “These questions address not only what brings customers to your company but also what makes them leave," said Lucker. "To begin to answer a question like this, you have to understand the company's product strategy, the various segments of customers that are served, what drives people to buy the products, what it is that 'closes' the sales deals with people, and also the various signals along the way that people leave. An analysis of all of this ultimately leads you to what you think you can do to elicit the kinds of behavior that you want."

While a well-written piece, I do think Lucker is selling his clients short.  In my experience, business managers are not only asking these questions,  they also know what to do with the answers.  The big challenge most of them face is having the time to correlate the individual indicators that Lucker describes in his example to what is going on in the business.

Today’s business intelligence and reporting-driven analysis environments focus on delivering results from individual tests. Once the initial analysis is complete, it is up to the business manager to take these results, analyze them, reach a conclusion, and determine the appropriate next course of action. This is similar to the way doctors review the panel of tests resulting from a blood test.  It is up to the individual physician to review, correlate, and analyze the results based on an individual patient’s history. It sounds exhausting, and quite simply, it is.

The challenge is not the physician’s ability to determine the best course of action or ask the right questions.  The challenge is in the time required to review, correlate, and analyze the individual results in the context of the patient’s history.  Further, the opportunity for error in this manual process stems from the number of individual results combined with the physician’s bias in looking for certain markers There is also the added pressure to make a determination quickly so that he or she can move to the next patient’s results.

While business people don’t play with the “life and death” stakes of a physician, there is still pressure to balance efficiency with the amount of work while keeping resources to a minimum. That’s exhausting too!

The Big Data ecosystem is obsessed with data analysis meeting the definition of “Big Data”, i.e., satisfying the “three Vs”: volume, variety, and velocity.  The Big Data ecosystem is also overly consumed with definitions and determining whether problems are “big enough” for Big Data.  I think this approach is incorrect, as the reality is that value is determined by the ability of data analysis to help companies manage businesses better and make better decisions. Often, this means bringing together data elements that are not normally associated with one another.  It means automatically correlating results so that answers are delivered instead of clues being amassed.  And it means answering questions that can spur action, not just interesting information.

Is your data “big enough” for big data? I think the answer is always “yes.”

At Oversight, we focus on automatically analyzing 100% of our customers’ transactions so they can see the behavior that is represented in those transactions.  Understanding behavior and the underlying causes for that behavior drives action and improvements.  Sometimes the data involved represents a lot of volume, often there is a lot of variety, and the velocity (or timing for action) varies greatly.  The Big Data ecosystem will argue that our Insights On Demand solution doesn’t qualify as “Big Data” because one or more of the “three Vs” is missing, or not big enough.  That's fine by me as long as our customers agree that Insights On Demand is providing "Big Answers".

But the reality is that whether the data analyzed is “big” or not, “Crunchy” or not, effective data analysis is about driving value in the business.

spend-analysis-vol3