Follow Datanami:
July 24, 2013

Attention to Small Details a Key to Big Data Success

Alex Woodie

Companies may spend millions of dollars assembling a team of data scientists and analytic tools to process huge amounts of structured and unstructured data. But when it comes to improving processes, products, or profits, a laser focus on small details is often what counts the most in big data.

ComputerWorld recently published an illuminating look into the big data projects of three Fortune 100 firms, including Intel, UPS, and Express Scripts. All three companies have been able to save money and improve their offerings by using big data analytics to uncover aspects of their businesses that may have otherwise gone overlooked.

For example, pharmacy claims processor Express Scripts is able to save its customers money by detecting whether or not they are utilizing the lowest-cost fulfillment option for their prescriptions, which is often mail-order fulfillment from the company’s pharmacy. When the company finds a customer who is overpaying, they will send out emails to encourage the customer to visit the company’s website to change their preference and begin mail-order delivery.

It boils down to “doing the data analysis, creating the interaction, and getting out the right message so that the patient can make a different choice,” Express Scripts CTO Jim Lammers told ComputerWorld.

At United Parcel Service, sensor data from delivery trucks accumulates in On-Road Integrated Optimization and Navigation (ORION), a big-data platform designed to help build efficiencies into routing and preventative maintenance. By grouping routes into loops that famously avoid left-hand turns, ORION has helped eliminate millions of miles from UPS routes, and saved millions of gallons of fuel and reduced carbon emissions in the process.

ORION can also peer into little things, such how often a driver backs up or makes a U-turn. These time- and fuel-wasting maneuvers indicate a poorly trained driver. ORION identifies these drivers to UPS, so the drivers can be given addition training.

“We have sensors that capture information about the vehicle and the driver’s behaviors. We marry that information to delivery and acquisition information, and we can get a complete picture of how a driver is completing his work, day in and day out,” Juan Perez, vice president of information services, told ComputerWorld. “That has incredible consequences for the way we manage the business across the board.”

Chip giant Intel has several big data projects in the works. One of the projects involves using predictive analytics to accelerate microprocessor testing time. During a proof of concept project, the analytic software saved Intel about $3 million, a number that could grow to $30 million as its implemented more widely next year, according to ComputerWorld.

Other projects are tackled according to the “six months and $10 million” rule. These projects–which last six months and aim to save the company $10 million–are completed by teams of five people, each of whom has a different core skill, such as business expertise, statistics, predictive modeling, machine learning, and data science.

“Each person on the team had a slightly different perspective on the problem we were trying to solve,” Intel CIO Kim Stevenson told ComputerWorld. “Doing it in six months was our way of earning the right to prove the capability was there to really change the way we do things.” Intel has completed 13 such projects, and is now working on projects with $100 million payoffs.

Related Items:

DataStax Rakes $45 Million; Schemes Growth

The Three Spokes of Wisdom in the Wheel of (Big Data) Life

Nate Silver on the Art of Statistical Analysis