[Editor's note: This article is by Michelle Zhou, a senior manager at IBM Research – Almaden. Dr. Zhou will participate in a panel titled "Big Data in the Entertainment Industry" at the IBM Research Colloquia "Box Office to Front Office: Winning with Big Data" on Aug. 10, 2012. Watch over livestream beginning at 10:00 a.m. US Pacific Time.]
My husband and I like to go to our local library and will occasionally borrow movies. But without a way to search beyond general categories in alphabetical lists, it's hard to find something we want to watch. With Netflix, on the other hand, we get recommendations based on our past rentals and movies we've rated.
My team at IBM works in people analytics, which aims to gain a deep understanding of people, including understanding their personalities and needs. We then use this understanding to create hyper-personalized engagements with individuals (e.g., making a movie recommendation based on one's personality). Right now, we're looking at how to use big data generated on social media to analyze brand perception and the people who voice those perceptions. We want to help companies better serve their customers with what they want, when they want it.
Our work will also help companies better serve new customers who don't have enough behavioral data for the companies to analyze. Social media data can help companies find people with similar interests to those they already serve. In addition, social media helps companies learn about what their competitors' customers are like and how they are served.
I get an average of 50 email campaigns in my personal email inbox every day. Out of those, I'm lucky if I find one that has something I need at that moment. But I don't unsubscribe to the rest, because I might need something from one of them in the future.
At IBM, we're trying to figure out how companies can send these campaigns only when you want them.
The Four Vs
IBM defines big data according to volume (the scale of the data), velocity (how fast the data is moving), veracity (how accurate/truthful is the data), and variety (forms of the data).
Today, not only is traditional data (books and sales transactions) captured, digitized, and made more accessible, but new types of data (social media and mobile activities) are also being generated at an incredible pace.
IBM has helped its customers manage their data in an efficient, effective, and secure manner -- even before the big data boom. Now, our analytics software and services can help them extract more value from all the data they own.
Discovering insights to help both businesses and individuals
Some would say this is just another effort to sell more to more people. But I think that the better we understand our customers, and people in general, the better we can serve them and help them.
Analytics is more than targeted advertising. It can be a public service.
For example, after the 2011 earthquake and tsunami in Japan, a rumor spread through China that iodized salt would prevent radiation sickness. That false statement caused a rush on salt at supermarkets across the country, leading to shortages and to scalpers looking for quick profit. What if China's government could have detected and responded to the rumor as quickly as it spread?
Whether it's helping an individual find a product or service that he or she wants, when he or she wants it, or helping a government agency respond to a crisis, my team wants to find ways to use big data (volume) in real-time (velocity), which may come in different forms (variety), to give people the most accurate (veracity) and useful information possible.
Watch "Winning with Big Data"
On August 10, I will join a panel with Todd Yellin, the vice president of product innovation at Netflix and Ray Elias, the CMO at StubHub, as part of "Box Office to Front Office: Winning with Big Data."
Add a reminder to your calendar here to watch it live, and join a chat with a member of my team, Jeff Nichols, on IBM's Smarter Planet Facebook page.
— Michelle Zhou, senior manager at IBM Research – Almaden.