Earlier this year, leading job site GlassDoor identified the 50 best jobs in America in 2017, taking into account the earning potential based on median annual base salary, the rate of job satisfaction, and the number of job openings.
The job that garnered the number one spot? Data scientist.
That’s not surprising to Robin Opie, vice president of data science at Oracle. Opie runs a team of 140 data scientists and analysts who make the products that Oracle Data Cloud sells in the marketplace. As head of the analytics and data science team, Opie says half of his time is spent hiring, developing, and retaining the talent required to write software that can digest and glean business insights from high volumes of data.
“A lot of people think of data science as this one brand of math nerdy stuff,” says Opie. “In reality, there is tremendous diversity in the types of things you can do as a data analyst, and the work is incredibly rewarding on many levels.”
So what does it take to get the best job in America, and what makes it so great? A closer look at Opie’s team at Oracle Data Cloud provides a wealth of information. Necessary traits include curiosity, a passion for problem-solving, and a surprising lack of ego.
Oracle Data Cloud helps businesses better target prospective customers with highly accurate demographic and consumer data. The data-as-a-service platform was built by unifying a number of industry-leading data platforms—including Datalogix, BlueKai, AddThis, Moat, and Crosswise—which aggregate information from online and offline sources.
Oracle’s data cloud analytics and data science team works on two primary machine learning products: one that produces data-driven audiences for client advertisements and one that measures how well those ads perform. Using de-identified data elements tied to $3 trillion in consumer transactions and de-identified demographic profile information on almost every household in the United States, the product can identify and target very specific groups of people for an advertiser.
For example, a women’s apparel manufacturer looking to run a digital ad campaign during the summer would want to target people who already buy high-end women’s apparel. Using their software, the team members can rank the consumers who match that segment. But with data and methods, the team can go one step further—it can give the manufacturer cookies and mobile IDs associated with the consumers who are most likely to buy the most women’s apparel this summer.
Then, after the ad has run, the team uses the measurement software to show advertisers how much more of their product was sold because of the media dollars just spent.
“As an example, an advertiser might use our audience targeting to reach 40% of summer women’s high-end apparel buyers, versus the 10% they would have reached without targeting,” Opie says. “This is a more efficient way of spending your valuable ad dollars.”
The analytics and data science team works directly with advertisers to identify audiences on platforms such as Facebook, Twitter, or the open web. It also sells audience information on the BlueKai Exchange, which is the world’s largest third-party data marketplace and the standard for open and transparent audience data trading. It provides an ecosystem built on premium quality data, flexible and fair pricing, and scale that is unmatched in the industry. “People can go buy our soccer mom segment, our heavy salty snack buyer segment, or whatever they may be looking for,” Opie says.
What It Takes
Crunching all of this data requires deep knowledge in fields such as statistics and computational science. For the team’s entry-level positions, Opie looks for recent graduates from grad programs in computer sciences or applied math disciplines who demonstrate a drive for reaching outcomes to business problems. He also looks, surprisingly, for “low ego.”
“We’re not just looking for people that are going to contribute a new approach to solve a problem,” Opie says. “We need people who can be part of a very collaborative team, and who are lifelong learners. If someone thinks they already know everything, it conflicts with those traits.”
Opie also looks for people who demonstrate passion—both for their work and their interests outside of work—and people who respect data and are willing to stand up for it, “no matter how hard the truth about what it says.”
Zach Knowlton is a data analyst on the team. His master’s thesis was on developing a way to measure the impact different players and positions have on scoring in college football. “Business acumen and coding skills can be acquired,” Knowlton says. “However, people who are naturally inquisitive and have an internal drive to follow through on solving tough problems tend to have great success in our space.”
Alex Sadovsky, director of data science at Oracle Data Cloud, says the best data analysts are able to solve problems using data-driven storytelling. As a student, Sadovsky wanted to understand how the brain is wired to form complex computational networks. After obtaining a PhD in computational neuroscience at the University of Chicago, he joined Datalogix, which Oracle acquired in 2015.
While he says it might seem crazy for someone with an advanced degree to shift from biological science to marketing, Sadovsky believes many of the problems translate across both practices. “The same approach I was using to understand which genes affect blood pressure I applied to figure out how consumer behaviors could lead to a purchase. Similarly, the work I did to understand how individual cells in the brain are connected to one another translated to understanding how email addresses, web browsers, and mobile phone apps link to an individual consumer in a physical household.”
A Day in the Life
A typical day for Knowlton involves working with the pipeline of marketing campaigns that need be measured. He customizes parts of the measurement product to fit specific campaigns and spends time debugging and modifying data. He also works on research projects to enhance the team’s products. “There are so many different pieces within our campaign process, so every day has its own nuances and challenges,” Knowlton says. “Which keeps work interesting and fun.”
Sadovsky runs the team responsible for developing the algorithms and software that produce Oracle Data Cloud’s audience products. His experience on the team has shown him that there is no one-size-fits-all method to modeling consumer behavior because human beings are incredibly complex and unique. He points to the example of data gleaned from Datalogix, which centered mainly on offline purchases, versus AddThis and BlueKai, which were focused on online data. Sadovsky says that one of the most interesting things he encounters every day is how data from online behavior differs from offline purchasing.
“The internet is often an aspirational medium, while in-store purchasing is reflective of true intent. Combining the two gives us a unique full-circle view of the consumer, which would be lost if you cast a blind eye to either side,” Sadovsky says.
In that data environment, Sadovsky’s team tackles a dizzying array of data questions, from “Who in the United States owns a dog?” to “Which visitors to a website are really planning to buy an expensive sports car?” He says one of the most rewarding parts of his job is enabling clients to find the answers to these questions in their own data. “Nothing is better than being able to help a client reach a eureka moment, where they realize that their incentives of improving their marketing approaches align perfectly with the data science solutions my team creates,” he says.
Challenge as Reward
While competitive compensation is fundamental to attracting and retaining such high-caliber data scientists, Opie places great importance on keeping them happy and interested.
“I think what’s as important for this group is finding new and important problems for them to solve,” Opie says. “In fact, we often joke about it—the reward here for a job well done is that you just get a more difficult job. But it’s something this group of people embraces.”
New members of Oracle’s data cloud analytics and data science team also have the opportunity to participate in a Rotational Analyst program, which exposes them to six months each of four different analytics disciplines in two years. The disciplines are in the fields of research and development, client analytics, delivery strategy, and custom analytics. At the end of two years, analysts are placed in the department that is the most fun for them. “So, for instance, the people who are hardcore researchers, looking for new theoretical constructs to answer hard questions, end up in R&D. But they bring with them the knowledge that they gathered in all of the other departments,” Opie says. “They also become mentors for other folks.”
Knowlton says the program has given him the ability to develop a wide range of skills. “Getting together with very smart people and observing how they think about a problem, then coming up with a group solution or multiple potential solutions, is always rewarding for me,” he says. “While you have to start over with a new team every six months, the experience and growth is worth it.”
Oracle Data Cloud’s analytics and data science team has grown rapidly and has had very high retention rates. Opie is currently looking to add 15 to 20 more analysts to the team.
“Essentially, when we find somebody who is fantastic, we make room for them,” Opie says. “If there’s a data scientist out there who loves these types of tools and wants to see how they make an impact, we can give them a great place to work—and with some pretty great people, to boot.”
Monica Mehta is a frequent contributor to Profit, Oracle’s quarterly journal of business and technology.