Tampa International Airport is playing its own version of Moneyball – and winning – as it has turned to data analytics to help it compete more effectively against larger airports in Miami and Orlando.
The airport uses analytics and data mining to uncover evidence that a particular airline could make money by offering flights to and from the Tampa Airport. Then it takes that data to the airline to convince it to add flights.
The effort has paid off as data analytics has helped the airport secure new daily routes to Cuba and Switzerland this year, boosting international travel by 20% in the first seven months of 2012.
“We’re a culture that’s completely data-driven,” Chris Minner, the airport’s vice president of marketing, told the Tampa Bay Times. “We were up every night until 1 a.m. burning through the airline data, trying to crunch the numbers and figure out where are the missed opportunities for Tampa International? What are the top underserved markets?”
But while the Florida airport has honed in on a crucial business problem to tackle with big data analytics, many executives are struggling to find a business use for the massive amounts of data they are collecting, say several IT executives in a recent Computerworld article.
Reid Nuttall, CIO of OGE Energy, an Oklahoma City-based energy company, tells the magazine that he believes that the large amount of data it amasses from smart meters that span its customer base will help OGE analyze and influence customer behavior in the future.
He’s looking for people within the company to start unearthing that insight from the data.
“Big data is forcing IT and business intelligence [teams] together” so they can work together to find ways to explore new data, he says.
William Herridge, managing director of emerging solutions at the Tribune Company, compares the big data challenge to the struggle to get business users to buy into the value of online analytical processing (OLAP) tools many years ago.
“We see the value in this, but getting users to understand that value” is challenging, especially when dealing with unstructured data that is inherent in big data, Herridge tells Computerworld. “Until business users can see some benefits, they are not going to sign on to big data projects.”
To help senior managers jumpstart big data projects, Harvard Business Review has posted three questions management can ask analytics teams to answer:
1. How will they coordinate multi-channel data? Companies must decide how they can create ways to easily merge data from brick-and-mortar stores, online retailing, social media and mobile. One option is to create a cross-channel identifier.
“Ultimately, that’s the value of a common identifier for any business: a fuller picture of related data under a single listing,” HBR notes. “[In retail] a single registration account for web and mobile commerce can help consolidate data from both channels in order to give a better picture of a customer’s online shopping. Even more broadly, a customer loyalty program can help, since it gives consumers a unique ID that they apply to every purchase, regardless of the channel.”
2. How will they manage unstructured data? Some database systems allow room for fields, comments or attachments that often have valuable data that can be attached to records. Or, a company could explore using metadata – data about data – by tagging database records with descriptions to help assemble them into categories.
3. How will they create the data needed to inform better business decisions from the data now on hand? For example, many auto insurance companies are beginning to use programs that track driving patterns through in-car devices. This data is entered into predictive models to determine potential driver risk and to determine premiums.
“The idea is that direct driving behavior over time will be more predictive than traditional proxies such as age, credit rating, or geography,” HBR notes. “While this seems like a logical assumption, the real question isn’t whether driving behavior is more predictive than traditional proxies – but whether driving behavior combined with traditional proxies is most predictive of all. What’s the equivalent of a driving score for your organization?”
Next Steps:
- Subscribe to our blog to stay up to date on the latest insights and trends in data analytics and big data.
- Register for the Spotfire 5 Webcast with Lou Jordano (@loujordano) on Wednesday, Oct. 17 at 1 p.m. Eastern.
