Join the local supermarket’s loyalty program and receive personalized discount offers on popular brands. Subscribe to Netflix, and it will send suggestions of the next movie you should watch. Add your business card to a fishbowl at a favorite restaurant for a chance to win a free lunch or dinner.
To an average consumer, these are clear-cut examples of how sharing data provides benefits. And to the companies utilizing big data, they are illustrations of how it helps them retain and deepen existing customer relationships and develop new ones.
Yet these simple, everyday uses of data merely scratch the surface of how the increasing volume of data—both structured (e.g., customers’ mailing addresses) and unstructured (e.g., social media streams)—impacts our lives. More often than not, we are unaware of how much is collected, how it is collected, and how it is used.
Certainly, news reports about massive data breaches and misuses catch our full attention. While stories about lax data privacy policies at Facebook and data leaks at Marriott Internationals’ Starwood Hotels are in the headlines, we maintain a heightened awareness of the power of data and begin to wonder if we should delete all of our social network profiles. When the hype dissipates, however, we return to posting to Facebook, Instagram, Pinterest and SnapChat freely, filling out customer experience surveys for retailers we frequent and entering our interests and exercise and sleeping habits into our fitness apps.
On the receiving end of all of this information are companies and organizations that understand the importance of data. Indeed, the value of data cannot be underestimated. The Economist regards data as the world’s most valuable resource—even more so than oil. While the latter heats homes and is a critical element to a wide variety of products, including medicines, it cannot, like data, help companies achieve specific business objectives—goals that are not as obvious as encouraging return visits to a grocery store, and—maybe—even a bit surprising.
Data: A strategic asset
Companies with a data-driven culture handle data with the same level of respect and regard they give to more commonly thought-of assets, such as trade secrets and equipment, in order to ensure the information delivers value. As such, they entrust data scientists to work with company leadership to develop and execute an enterprise-wide analytics strategy that aligns with the companies’ overarching goals.
Defining a plan is critical; without a roadmap, companies can fall victim to what is known as the “big data buzz”—collecting as much data as possible without a clear understanding of its quality or how to use it.
"One of the biggest problems is making data usable." – David Fusari '89
John Gillooly ’07, Managing Director, Data Analytics at Athena Global Advisors, is committed to helping clients avoid those pitfalls. A strategic consultant focused on assisting stakeholders make use of data-driven insights, Gillooly emphasizes the importance of focusing on goals.
“My first order of business is to figure out what my clients’ objectives are,” says Gillooly. “They may have a broad concept of what they want, but I have to dig deeper because the data they ask for is not necessarily what they need.”
“Once the goals are defined, we can then start talking about the data needed to support the goals, the data collection methods and how the data will be used.”
With strategies in place, it also incumbent upon data professionals to determine data quality, maintain data hygiene and sort it properly. Data can only become a working asset if it is error-free and arranged in a meaningful order. These are challenging tasks for both in-house and outsourced data teams given the inconsistencies in large data sets and the increasing amount of new data created—especially in ever-changing industries.
A Bryant alumnus with first-hand knowledge of the complexities involved in carrying out these responsibilities is David Fusari ’89, former Chief Technology Officer at TriNetX. The Cambridge, Mass.-based firm combines longitudinal clinical data—data derived from repeated observations of the same populations over a period of time—with state-of-the-art analytics on a self-service platform.
“One of the biggest challenges is making the data usable,” says Fusari. “Patient records from hospitals and other health care organizations often contain only one small, valuable element. I saw records referencing pregnant males or bizarre dates of birth. The general rule is to exclude that data.
“Medical coding and the rapid rollout of new drugs further complicate the process of curating the data. ICD codes (International Classification of Diseases) change all the time, so normalizing the data is a constant.”
TriNetX also uses natural language processing (NLP) to optimize the value it receives. For example, electronic medical records contain structured data, like the ICD codes, as well as unstructured data, such as written clinical notes. Using NLP allows TriNetX to extract data elements such as tumor sizes and characteristics from this text, map them to standard clinical terms and make them available to researchers.
Data: A collective force for good
Data—thoughtfully organized and cared-for—as it is at TriNetX, then provides organizations with opportunities to meet their goals that were seemingly unattainable prior to the data revolution. As an illustration, Fusari discussed the problems pharmaceutical companies had in identifying patients who qualified for clinical trials.
“Bringing a drug to market takes many years and costs about $2 billion to $3 billion with a large percentage of that spent [on R&D] before the drug is even tested on a human. All too often, the pharmaceutical companies would receive the FDA approval to move forward with clinical trials only to determine that there is a lack of patients who qualify for the trial.”
TriNetX helps mitigate this costly dilemma with its analytics platform that gives its customers access to tens of millions of anonymized patient records. Having on-demand access to this volume of clinical and claims data—representing more than 300 million patients—enables pharma companies to identify clinical trial sites more easily.
Other users of the TriNetX platform include clinical researchers working either for one hospital or collaboratively with peers across a healthcare network that shares data with the company. Researchers looking to identify more effective treatments for specific diseases or who are analyzing trends among specific patient populations can perform observational studies in real-time.
"Employers want to know what is going on in their population." – Janvi Nerurkar '14
Identifying trends through claims data is also of particular importance to U.S. employers that want a deeper understanding of their employees’ well-being. Using claims data analysis that insurers provide, companies utilize the information to make data-driven decisions about a host of issues, such as wellness programs, health benefit plans and ergonomically correct work environments.
Among the many analysts sifting through these immense data sets is Janvi Nerurkar ’14, Data Analyst II at Marsh & McLennan Agency, the ninth largest insurance broker in the United States. Part of the Planning & Analytics for Total Health (PATH) Department, Nerurkar aggregates and evaluates HIPAA-compliant insurance claims data to provide employers with a “health snapshot” of its workforce.
“Employers want to know what’s going on in their population,” says Nerurkar. “An analysis can indicate a prevalence of specific health issues like diabetes or smoking-related problems. In cases like these, the carriers often encourage employers to promote insurer-provided services like smoking cessation programs. If the carriers don’t have such programs in place, then our health management consultants can recommend vendors.”
Marsh analysts also compare the findings of peer companies of similar size in the same industry. By generating this benchmarking data, Marsh provides its clients with many benefits, including the ability to identify common claim types within their industry and the opportunity to develop practices that prevent the occurrence of those claims.
Data: A sales propeller
Just as TriNetX and Marsh & McLellan are sorting and analyzing data for organizations other than themselves, a broad spectrum of companies is engaging with big data to drive their business forward—generate more revenue, increase profits, and gain a stronger foothold in their respective markets.
Gillooly described how Timberland, the outdoor shoe and apparel company, collected data (unbeknownst to its customers) from social conversations and used it to refine its marketing practices and, ultimately, strengthen its brand equity.
“Say Timberland in the U.S., and outdoor activities, like hiking, immediately come to mind,” says Gillooly. “Yet, conversations on social media centered around how cool Beyoncé and Jay-Z look in their Timberlands. The brand name was frequently mentioned on hip-hop blogs and other online forums. Other discussions revealed that in Europe, Timberland is fashionable.”
These conversations, coupled with a multi-year customer study, prompted Timberland to shift its marketing and merchandising strategies within the last five years. While its products are still popular with outdoor enthusiasts, Timberland boots and apparel can be found not only at REI but also at Macy’s and Nordstrom. The changes have proven to be successful. Timberland is ranked among VF Corporation’s top five brands. (VF Corporation owns the Timberland brand.)
Timberland’s transition to a more fashion-forward company was made possible with voice of customer data; people were openly sharing their experiences with and expectations of the brand in very public forums.
"We look at hundreds of thousands of open deals and identify which may fall through." – Matt Los Kamp '13
In a B2B environment, however, it’s a different story. Given that the powers that be at Fortune 500 companies aren’t tweeting about their purchasing practices or intentions, predicting the probability of sales pipeline conversion becomes an exercise in data analytics.
Heading such an exercise at Dell Technologies is Senior Data Scientist and Team Leader Matt Los Kamp ’13. Using the Pipeline Health Index (PHI), an in-house developed analytics model, he and a team of analysts assess sales pipeline risks based on various data elements.
“We look at hundreds of thousands of open deals and identify which may fall through,” says Los Kamp. “Our predictive model considers many factors, such as the company’s annual revenues and its purchasing history with Dell. Each deal receives a PHI score; all assessments are shared with sales and finance executives across the enterprise.”
In addition to helping the sales teams close their deals, the data analyses also contribute to their ability to sell more effectively. By revealing commonalities among high-risk deals, the analytics provide the sales force with the opportunity to identify and mitigate those risks early on in the sales cycle.
Data: The ultimate multi-tasker
Tim Berners-Lee, inventor of the World Wide Web, once said, “Data is a precious thing.” Truer words were never spoken. The proliferation of data can help companies remain competitive, drive medical advances and improve individuals’ well-being. It is difficult to think of any other tool that provides as much meaningful information for making better decisions.
Certainly, there are drawbacks associated with data. When it’s misused and misinterpreted, consumer privacy is compromised, and decisions are made based on unsubstantiated facts. For these reasons, it’s best to remember what Kate Crawford, co-founder of the AI Now Institute at New York University, is fond of saying, “With big data comes big responsibilities.”
This story appears in the Summer 2019 issue of Bryant magazine.