At MIT Class for Executives, a Discussion of Big Data ‘Measurement Revolution’

by   |   November 7, 2012 2:06 pm   |   0 Comments

CAMBRIDGE, Mass.—Nearly 150 executives recently gathered in-person and online for a two-day course, a kind of Big Data 101, at MIT Sloan School of Management to learn about the opportunities and challenges of putting critical data assets at the center of their strategies.

“Big Data-Making Complex Things Simpler” included research on how manufacturers now view, in immense detail, how customers use their goods, said Prof. Erik Brynjolfsson, director of the MIT Center for Digital Business. That pays off for both sides of a transaction, large or small.

Related Stories

Business problems suited to big data analytics.

       Read more»

Sprint, N.Y. Stock Exchange look to turn analytics expertise into data products, services.

       Read more»

Improving the effectiveness of customer sentiment analysis.

       Read more»

For example, sensors in Volvo trucks gather 1 gigabyte of data per hour and detect vehicle usage patterns and track maintenance conditions (such as driver speeds, hours driven, and measures of engine and brake wear). Collecting this data enables the company to resolve problems faster and with better precision, leading to improved management practices. Aggregate repair data and future sales forecasts are based on actual vehicle use—not past sales reports—and yield data about feature sets based on experience, not what focus groups report was important.

Data Enabling ‘Measurement Revolution’
Volvo’s experience is a symbol for what Brynjolfsson describes as a “measurement revolution,” a change as important as the invention of Anton Van Leeuwenhoek’s microscope which enabled microbiologists to see organisms that were only speculated about in previous generations. Data trails will deliver new insights on how work is done within, and across, an organization, creating greater transparency.

In addition to analyzing their own operations, companies are getting fresh approaches from data-driven outsiders, Brynjolfsson said.

Tools such as expert networks and competitions open to the public provide great value at low costs, as when Allstate Corp. opened a challenge project with Kaggle, a San Francisco company that runs prediction modeling contests for data scientists.

Allstate used claims data history and improved by 271 percent the predictive rates of bodily injury based on auto characteristics. New insights came from old data—and adding proprietary details of location or individual client driving records meant local risk management could be customized, instead of averaged against the much larger data set.

Social network analysis tools that spot patterns in the frequency and spread of an email or new phrase on Twitter or Facebook represent another measurement channel, said Prof. Alex “Sandy” Pentland, director of MIT Media Lab entrepreneurship program.

For example, advertising companies can create models to measure the spread of a new idea and to see if the uptake data matches predictions. Simulations can examine social channels through online, trackable data such as cell phone location and email networks.

These social networks are physical as well as digital, Pentland said. A person’s entire network, for example, represents opportunities to solve problems, change behaviors and reach new unexpected audiences beyond one individual. This is especially true in the workplace, where influence is greatest among the people we spend the most time with. In a project for Nokia, for example, Pentland said he found that consumers most often bought handsets they had seen in use around them, rather than models advertised.

Consumer’s Privacy and Big Data
Attendees at the course also heard about consumer data privacy concerns.  “Regulators are realizing that personal data is like cash – it can be stolen, falsified, copied, bought and sold,” Pentland said.

Pentland cited mobile telephony data as a realm where valuable user data is simultaneously claimed by device manufacturers, carriers, customers, and applications makers.

Pentland predicts a global trust network or clearinghouse of digital identity will emerge from the need for both safeguarding personal data, while trading on details within for the end-user’s convenience.

One experiment worth watching, he said, will occur in Trento, Italy where a consortium will harness public and selected private data securely to streamline government and personal services while giving individuals and organizations control of the data they create.

The course was designed for a general executive audience. And while the event provided useful insights into current trends, it was not about teaching how to manage big data strategy, said Aamir Mohiuddin, CIO of AmeriStar Information Network LLC, a Dallas-based real estate information company.

“Though the analyses were good, exposure to more current analytical tools and technologies could have provided a better understanding of the subject and its applicability in a rapidly-changing business environment,” Mohiuddin said.

David J. Wallace, a freelance writer and communications consultant, has written for The New York Times and Knowledge@Wharton. Follow him on Twitter: @gamechange.

Tags: , ,

Post a Comment

Your email is never published nor shared. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>