Chris Anderson (follow @chr1sa) of Wired Magazine in his opening day keynote challenged us all to embrace both our inner quant and the astonishing power of big data that drives our businesses today.
“We are trained to think about the average,” he said. Yet, that focus on what “most people do” is a very 20th Century way of thinking about data and the customer. However, today we market based on interest and preference and behavior, not demographic or biology (like shoe size or gender).
“When it’s something that is characterized by our intellect, choices, mind, preferences – then “most” is not applicable,” Chris said. This 21st Century approach recognizes that people are defined by their unique qualities rather than one size fits all. Big Data allows us to avoid the “middle” – which is the wrong spot for the digital age. The long tail is a better model for understanding the broad spectrum of our customer base and marketplaces. “There is a small group of things that is very popular and a lot of things that are not,” he said. However, add up all those “not popular” and you get a big market.
Chris urged us to recognize that targeting people’s niche interests is a good way to capture market share. This is big data thinking. Welcome to the world of the quants.
Chris offered some very mind bending ideas – for more, grab a copy of his new book, “Makers – the New Industrial Revolution” which is available here at the show. Of course, his classic, “The Long Tail” is always worth a re-read.
Consider these morsels of wisdom:
- The 17th Century economic Thomas Bayes still drives a lot of statistical thinking. Basically, the Bayesian theorem is that you can’t know for sure. There is no right, just a righter. Probability not certainty.
- What questions are we asking of the data? Every data set offers opportunity to find connections. The amount of data increases exponentially so do the number of hypotheses. You can come up with answers even if you don’t know why.
- Of all the infinite number of correlations between the data, which ones to act upon? The answer is to use your common sense. This requires a “fundamental intellectual leap” as Chris calls it, to apply your own skills and insight to construct experiments on the data.
- A/B testing is a good way to use big data. You don’t have to predict, you just test ideas until you can build on evidence rather than guesses. However, Chris warns that A/B testing alone will not help you produce big insights. You also need visionaries and quants to help you conceive of new correlations and causations and to try something radically new. “Once you get into a new method, you can use A/B testing to refine it,” he said.
-Stephanie Miller, VP, Member Relations, The DMA