Jim Sterne on Big Data and the Human Side of Analytics
Jim Sterne, Founder of the eMetrics Summit and co-founder and current Chairman of the Digital Analytics Association will be one of the Keynote speakers at Affiliate Management Days (AM Days) taking place April 16-17, 2013 in San Francisco.
Jim’s keynote, The Human Side of Analytics, will focus on social, mobile, affiliate, online, offline, and integrated marketing and how it all adds up to Big Data. But his main focus for the talk will be the art of analysis. For merchants and affiliate managers understanding how you go about asking the questions and solving the problems using the enormous amount of available data is crucial.
The PMA had the opportunity to ask Jim, a consultant who has written seven books on using the Internet for marketing and produces several conferences, about the rise of Big Data, how organizations deal with data, asking the right questions, looking for truth in data and more.
Performance Marketing Association: Big Data has been a huge buzzword for months. Why is Big Data suddenly all the rage? Or is this more of an overnight success that took the rest of us a decade to recognize?
Jim Sterne: Big Data is indeed an overnight success after many years. What made it pop out into the open? The advent of a LOT of data happened when we crossed over from tracking just transactional data to tracking behavioral (Web analytics) and attitudinal (social media metrics). The huge amount of data became ginormous. But that was just a problem. The solution came from the open sourcing of the technology that could manage large numbers of individual processors (MapReduce/Hadoop). At this point, the industry needed to call it something and “Big Data” stuck.
PMA: You are all about the data, but not everyone has a head for making sense of analytics. Where should the data averse start?
JS: That’s a little like saying, “I hate spelling,” or “I hate math.” You don’t have to actually do math and we all have spell-check now. There are many roles involved in analytics. Those who love the technology focus on capturing the data. Those who love the data focus on cleaning it and normalizing it and putting it together with other data streams. Those who love the puzzle focus on asking questions of the data to see if they can derive insights. Some people love the formulae that those analysts require in order to ask their questions and they become data scientists. The rest of us focus on the decisions we can make based on the insights the analysts come up with. So, while I no longer need to crunch the numbers myself, I still need to understand the basics so that I can tell if the waiter added up the bill correctly.
PMA: People often say that numbers can be manipulated to support any argument. Do you agree with that? If not, how do you insure that data reveals truths?
JS: Yes, numbers can be tortured to say anything. The way you ensure data reveals truth is to ask it questions rather than ask it to confirm an opinion. Andrew Lang (http://en.wikipedia.org/wiki/Andrew_Lang) cautioned that we “should not use statistics as a drunken man uses lamp-posts – for support rather than for illumination.” To guard against that, we formalize the hypothesis, test, analysis cycle and avoid that request for the ad-hoc report to prove the effort was worthwhile. “Show me some numbers that show we got ROI for this afternoon’s meeting!”
PMA: Analytics is part science and part art. Can you explain the art part of it?
JS: Sherlock Holmes was an artist. The science is collecting and cleaning and moving data around. The art is asking the right questions…and then asking the next question.
PMA: How can the average business person manage all the data they receive or collect – or at least not be overwhelmed by it?
JS: Get help. There are lots of consultants around and they can be very helpful.
PMA: How do managers go about determining the most important key performance metrics for their businesses?
JS: This is a question about corporate culture and politics. What you measure depends on your goals and your organization’s goals. There is no one-size-fits-all here. There are no universal KPI’s.
PMA: Should looking at the analytics and metrics be a part of each mangers job, or should companies think about having a separate analytics team?
JS: The pendulum swings at a different rate in each company. Analytics starts in one small area and eventually the people doing the work are asked to help others. Interest in analytics spreads from group to group until management realizes that some groups are doing it much better than others and consolidates them into a central center of excellence. At some point, this central group figures out how to embed analytics into each department to become a “data driven organization.” This works very well… for a time… until the processes stray from the original format and success, once again, becomes uneven. That’s when the pendulum starts to swing back to centralization.
Get the AM Days full agenda
PMA members will receive a 15 percent discount on registration. Check your PMA member-only newsletter for the promo code or email Rachel Guillot (firstname.lastname@example.org) for the PMA member-only discount.
Latest posts by Lisa Picarille (see all)
- Co-opetition: So When is it a Good Idea to Work With a Competitor? - November 9, 2015
- PMA Members Out in Full Force at ASE15 - August 17, 2015
- Let’s Meet at ASE15 - July 27, 2015