Business Science and Technology

The Data Machine

Value in the Era of Analytics
data, business miami, university of miami school of business
Art Credit: Dung Joang

We’ve all experienced the power of data analytics in action. It happens every time you tap into Uber to see how long you’ll need to wait for a car or scroll through the list of “Recommended for You” movies, books and products on Netflix or Amazon. It’s going on in the background when you apply for a mortgage, get quoted a premium for car insurance, or book a hotel room or plane ticket. Increasingly, businesses across industries are collecting and analyzing demographic, behavioral and contextual data to do everything from boost customer satisfaction and improve operating efficiency to streamline processes and choose the right job candidates.

At a time when both data collection and computing power are increasing exponentially, these examples of how companies are crunching numbers to unlock value are just the tip of the proverbial iceberg. According to a recent report by IBM, 90% of the data in the world today was created during the last two years alone. In fact, thanks to the growing prevalence of data collection methods – including internet search engines, cell phone signals, sensors, social media and downloadable government data – we now amass 2.5 quintillion bytes of data a day.

“For a long time, people were developing algorithms that they weren’t able to use because of computing limitations or lack of data,” says Daniel McGibney, assistant professor of professional practice in the School’s Department of Management Science and co-director of the School’s Master of Science in Business Analytics program. “Now that we have the necessary computing power, and that just about every industry either has data or can get it, you’re seeing data analytics taking off everywhere.” 

This untapped potential represents both an opportunity and a challenge for businesses, many of which are finding themselves inundated with information that they haven’t yet figured out how to mine, let alone monetize. “There is just a ton of data out there, and different companies and industries are at different levels in terms of understanding what they can do with it,” notes Doug Lehmann, assistant professor of professional practice in the School’s management science department and co-director of the Master of Science in Business Analytics program. “While a lot of people are still at the point where they want to jump in and do something with all the data they have without knowing exactly what, virtually every industry has examples of some standouts that are doing well in analytics.” 

Data analytics has already revolutionized business-to-consumer marketing, as more companies glean insights from information about potential and current customers and use it to tailor their messages to specific audiences. “Google, Facebook and other companies are collecting all of that information and either using it themselves to target ads more effectively or selling it,” Lehmann says. “Marketing is so much more sophisticated now than it was when you would just blindly send a message to 100,000 people and hope for the best. Now, you can accurately target the people who have the highest probability of following through with what you’re hoping to achieve and focus your efforts there.”

PROFITING FROM PROBABILITY

But the possibilities for consumer facing businesses stretch far beyond getting more bang per advertising dollar. The more progressive companies combine internal and external data and use predictive modeling to not only identify target customers, but to zero in on those who represent the most profit potential. They create complex models to do things like assess the likelihood of customer attrition and adjust prices when demand spikes or softens. Entire industries are being disrupted by startups based around businesses that have data analytics at their very core. 

At ride company Uber, for example, data on traffic patterns, passenger volume, weather and other variables churn through complex algorithms to predict demand, which then determines where its drivers should congregate and how much they should charge for a ride. The information driven model has proved devastating for taxi companies. They’ve responded with mobile booking apps, but they don’t change prices or reposition cars based on demand.

In a similar vein, the peer-to-peer lodging site Airbnb uses data analytics to optimize both the type of properties it targets for listings and the prices its participating hosts charge. In addition to variables like location, time of year and access to transportation, the company’s algorithms factor in the prices of rooms at competing hotels, which are losing business to travelers intrigued by Airbnb options. “Airbnb scoops all the hotel pricing information in an area and says, ‘Here’s the price you should set if you want your room to rent,’” explains Pete Gibson, founder of Datlytics, a data analytics consultancy that works with School of Business students who undertake projects for local companies as part of their management science coursework. “They have a huge data machine that the hotel brands don’t have.”

While industry disruption makes headlines, established companies making incremental advances lead most analytical endeavors. In industries like insurance, for example, data analytics is increasingly being used for a broad range of purposes. Analyzing claims and claim histories can help insurers predict the types of claims they’re likely to receive, as well as outcomes such as payout amounts and the likelihood of litigation. They’re able to use that information to optimize their limits for instant payouts and settlements. 

In sports, team franchises are looking at the demographics and purchasing patterns of fans to maximize ticket sales. “Everyone thinks about ‘Moneyball,’” says Lehmann, referring to the movie that showcased the use of analytics to evaluate potential players. “But there is a whole business side to analytics in professional sports. By looking at things like price elasticity and whether more people come on Saturdays than Wednesdays, we can consider the outcome of putting headline games on Wednesdays to draw a bigger crowd or the impact of a 10% price increase on a certain seating section.”

That’s exactly the kind of market research that Master of Science in Business Analytics student Kevin Keating has been working on as part of a project for the National Hockey League’s Florida Panthers. “We created a map of South Florida that pulls in census information for Dade, Broward and Palm Beach counties that we can sort by things like driving time to the stadium, age, median income and so on, to give the team a better understanding of their local market,” he explains. “Knowing the average income can help the team consider where a mail campaign might pay off or whether to offer a lower-tiered package in some areas and bump up prices in others.”

In health care, data analytics can both improve efficiency by helping physicians assess treatment options, and literally save lives. Image classification and processing technology have the potential to enable faster and more accurate analysis of CT and MRI scans to predict heart disease and diagnose health issues. Furthermore, DNA sequencing research is expected to unlock invaluable information about genetic characteristics that can be used to improve health care outcomes. “Researchers are basically trying to see what anomalies in your DNA are associated with various diseases,” Lehmann says. “If they can identify a certain gene mutation that’s associated with a specific condition, they may be able to develop a drug that specifically targets that gene.” 

UNDERSTANDING OPPORTUNITIES

The more you delve into data analytics, the more it becomes clear that the question isn’t whether data analytics can deliver value for a specific business or industry, but how to go about identifying and pursuing the opportunities it offers. “Car rental companies, for example, typically figure out how many cars they’ll need by looking at what they did last year and figuring that the market will go up by a given percentage,” Datlytics’s Gibson says. “But by using predictive analysis and [generating] demand curves, we can give them a much more accurate idea of the number of cars and type of cars they need, which can save a company millions of dollars across a fleet.” 

In working with companies seeking to mine their data to address a specific challenge, Gibson has found that projects tend to multiply as businesses get a taste of success. “We often find that people don’t know all the great things they can do with the data they have,” he says. “Once they start seeing the results, their eyes light up and they say, ‘Can you do this? And that?’ The other neat thing is that once the data has been prepared and the predictive model algorithms have been written, the heavy lifting is done and it can be relatively easy to pull more out.” 

As companies across industries recognize that leveraging the potential of data analytics is a competitive imperative, demand for people with an understanding of the capabilities of analytics and proficiency in data science is rising. “There’s an awareness that you need people who may not be data scientists, but who have the understanding to think about what analytics can offer their business,” Lehmann says. “They don’t necessarily need to be the PhDs in advanced math who create the tools, but they need the competency to be able to sit down at the computer with a data set and use the models to come up with answers for the questions.”

Today’s companies have an unprecedented ability to gather, aggregate and analyze data on production processes, sales, customer interactions and more. The potential exists to optimize virtually every major business decision – but companies need people capable of applying the kind of advanced modeling techniques that data scientists develop, McGibney agrees. “Strategic decisions are increasingly being made using data, so it is in a company’s best interest to hire people with the analytical skills to use these tools to inform their decisions,” he says. “People are increasingly at a major disadvantage if they don’t have a core understanding of business analytics.”

That thinking resonates with Keating, who enrolled in the master’s program determined to add a solid foundation in analytics to his repertoire. He saw a steady increase in the use of information technology during a career that spanned positions at several Wall Street financial institutions. “During my last two job searches, I noticed more companies looking for technical rather than financial skills,” he recounts. “I didn’t want to wind up having worked hard to achieve a level of success and have it ripped out from under me because I didn’t keep up with the skills you need today.”

IN THE CLASSROOM
Answering the Call for Analytics

With data analysis rapidly transforming entire industries, it’s increasingly critical to equip business school students with the foundational skills to understand how to use data to solve real-world problems. Launched three years ago, the School’s Master of Science in Business Analytics program is designed to do just that, says Doug Lehmann, assistant professor of professional practice in management science and co-director of the program. “It’s an accelerated, specialized program that enables people to get the skills necessary to understand, manage and apply data analysis in a business context in just 10 months,” he says. 

While undergraduates at the School can choose a major or a minor in business analytics, the master’s program is intended to enable students across disciplines to get up to speed in the latest data-related technology and tools. “We get a wide range of students in the program,” Lehmann says. “The common ground is that they all understand the importance of analytics.” 

The business analytics master’s program includes courses in areas like data modeling, data management, data mining and programming for analytics, and it’s updated annually to keep up with the pace of change in the field. “Next year, we’re adding a machine learning class and a generalized linear models class,” says Daniel McGibney, assistant professor of professional practice in management science and co-director of the program. He notes that, while jumping into sophisticated programming languages “is a bit of a learning curve,” students in the program have risen to the challenge and tend to progress quickly. 

In addition to months of coursework, the program includes a capstone – a real-world project that students complete for a sponsoring organization, under the oversight of an industry professional. 

The students receive a data set and a problem or situation to address; they are charged with finding and presenting actionable items or data-driven solutions to their sponsors. 

Past students have conducted projects for the Florida Panthers, Miami Heat and a local hospice. 

With demand for analytics expertise continuing to swell, Lehmann notes that the School is committed to preparing students for an increasingly datacentric business world. “We will continue to expand the curriculum to make sure students in all of the different disciplines are getting exposure to the analytics tools they need,” he says. “Employers are looking for these skills – not necessarily theoretical statisticians, but people who are able to tap into and harness the power of data in whatever it is they do.”

IN THE FIELD
Tapping into Analytics with the Florida Panthers

As any sports fan can tell you, data analytics plays an increasingly pivotal role in decision making at virtually every major professional league team. Managers and coaches who once made gut decisions about which players to draft, trade and develop – as well as about game strategy – now routinely pore over reports generated by mathematicians and factor in statistics when weighing their options. With more teams getting into the analytics act, it came as little surprise when the NHL’s Florida Panthers joined the race to leverage data by hiring Brian Macdonald as director of hockey analytics three years ago. 

Brought in to provide the team’s scouting, coaching and training staff with data that would inform decisions about team personnel, Macdonald soon found his role expanding to encompass business strategy analytics. “Our approach became: Anywhere that data exists, we want to analyze that data and use it to make smarter decisions,” he says. “And we had lots of data – from ticket sales, from finance, from marketing – to draw from on the business side.” 

Recently, Macdonald tapped into the School’s Master of Science in Business Analytics program, which assigns groups of students to work with local businesses on capstone projects that enable them to wrap up their studies with real-world experience. “We asked them to build us a data visualization tool that would give our marketing department a way to look at customer and census data on a map of southern Florida,” says Macdonald, who explains that the students were charged with acquiring, cleaning and organizing the data. “We didn’t come up with a project for the sake of coming up with a project; we asked for something that we really needed, wanted and will use going forward,” he adds. 

Kevin Keating, one of the master’s students who worked with Macdonald, says the experience of using the R programming language learned during his studies to work on a real-world project was invaluable. “We were able to build an application that enables the end-user to go in and select one of the criteria that we populated the map with – median income, for example – to see a color-coded map of South Florida by income,” he says. “We had exposure during the program to many different technologies – regression analysis, machine learning, high performance computing – and various programming languages. At the end, the work with the Panthers was an eye-opening experience in terms of showing us what pulling together information from disparate places can achieve.” 

The outcome was equally eye-opening for the Panthers, who have already gleaned market insights from the tool. It was a win-win. “The students got a lot of really good experience doing data analysis, and we got something that we are actually going to use,” says Macdonald.