Being Analytically Relevant
One of the people I met and enjoyed listening to at the TDWI Orlando conference last week was Piyanka Jain, former head of North America business analytics for PayPal. She now runs Aryng.com, a consultancy and service that trains and makes the ideas of analytics accessible to people with non-statistical backgrounds.
I decided right away that her service is greatly needed, equally if not more so for IT managers who attend data conferences to improve their knowledge of arcane databases, modeling and other details. If you're not making your living as a data scientist or theoretician, it's important to keep one eye on the prize of our handiwork with data, the problems we're solving and why. If you're not thinking this way, you're doing so at you own peril.
Though her service is meant to educate business types, and she has some high-end business clients, any analyst or data modeler worth their salt ought to have the same thoughts at their fingertips and practice what they preach.
Piyanka's simple mantra at TDWI had to do with the disconnect between our infatuation with data and the problem at hand. "Analytics is what sits between data and decisions," she told a sleepy early morning audience, and in business, enlightenment is supposed to lead to a reward.
At a conference with a theme of big data, Jain said it's just as important to be deductive as it is to contemplate new or huge sources of information. "With analytics, you can be an explorer or a detective," she said. Rather than start with the data, consider your approach to a business problem and that it's a suitable problem to attack.
In her example, imagine your job is to find gold lost in the Pacific Ocean. An analytic explorer, she said, could take an approach equivalent to building a submarine and scouring the seabed. Though great adventures and sights might await, it's a tough way to find gold and a poor match for your job description. Or, as she put it, imagine how excited your employers or shareholders would be with your anomalies and outliers.
An analytic detective, by contrast, would determine where the shipping lanes ran, the history of trade and depth of the seas. You could prioritize a few areas and stand a much higher chance of success.
She used this example as an entry to portfolio analysis, a measurement framework and customer analysis, that it's importance to know your measure of success and be sure your metrics point to that (and change if necessary).
Sometimes, it's the simple message that connect best. When it doesn't, that tells us something too. Looking around the conference hall, I could see her words sinking in with some people but not with others.
Well, it was an early morning session. But when it comes to relevance -- what are we talking about and why do we care -- I'm going to put my money on Sherlock Holmes.