Editor’s Note: This post features an interview with Yale University’s Dr. Trace Kershaw
Big Data remains a big buzz word, yet professionals struggle to understand how it can add real value. Over the next few months, I will be sharing ways to combine Big Data with traditional qualitative and quantitative research to help brands answer questions and motivate target audiences.
Today, we’re going to focus on how Big Data is helping researchers understand influence in social networks, leading to the design and testing of interventions that drive healthy behaviors. We’ll also talk about how this work can be extended to help brands that benefit from word of mouth and social media.
Health and the Influence of Social Networks
You can hardly open your newsfeed without seeing a headline lamenting the state of public health. Across a broad spectrum of disease states, much work lies in front of us to motivate healthy behavior in the US and across the world.
The good news is that we’ve never been in a better position to research, create and test programs that motivate healthy behaviors to promote wellness and minimize disease risk. Why? At least part of the answer lies in the Big Data stream of the Internet of Things: A stream of behavioral data that—when combined with traditional survey data—helps researchers capture, assess and build models of the social behavior associated with disease risk to design intervention strategies.
“Health can spread like a virus through a social network,” says Dr. Trace Kershaw of Yale University. “This finding was first demonstrated by Nicholas Christakis and James Fowler, whose work showed how individuals in social networks influence health. For example, they found that when an individual becomes obese, their friends are more likely to become obese, as are their friends’ friends.” This can occur through several mechanisms, such as:
- Direct influence: “Hey, do you want to go get a burger and a beer?”
- Social norms: “Bill eats a lot of steak and seems happy.”
- Social support: “You’ve had a rough day. Let’s go get some ice cream.”
Dr. Kershaw’s work builds on these findings to better understand how interpersonal relationships influence sexual, reproductive and maternal/paternal-child health. “We believe we can influence health through people’s social networks. If we can achieve a strong understanding of certain individuals, we can intervene and that can potentially spread through their network. The more we know about individuals and their network, the more targeted we can be in the intervention, prevention or cure strategies to most positively influence the health of the total group.”
At a high level, the research process looks like this:
- Recruitment: Dr. Kershaw starts by identifying individuals to participate in the research: “We work to understand these individuals so we know how to intervene at the right time to promote healthy behaviors.” They also recruit the individuals’ social networks in the hope that they will eventually spread healthy messages and positive behavioral modeling through their network.”
- Data collection: “The idea is to use as much data as possible to understand the problem and to create and test intervention strategies,” explains Dr. Kershaw. To accomplish this, they use an app to capture phone activity, including calls, text messages, emails, GPS coordinates and website visits. This Big Data set helps the research team understand the nature of information and how it is flowing through the social network. Traditional surveys are also administered to capture attitudes about health, risk and related topics.
- Analysis: Based on the data, the researchers will try to understand how the network communicates, build risk profiles and predict whether health risks will increase or decrease. He adds, “We have a hypothesis about which variables will be most important, but we’re not completely sure about which things will be useful. So, we’ll test different models to see if we can successfully intervene to minimize risk.”
According to Dr. Kershaw, the time is right for this type of research. “From a marketing standpoint, text messaging interventions have been going on for years. For example, if someone is near a retail store and they have opted-in to receive messages, the company can detect that they are close and send them a coupon to motivate them to shop. For instance, Amazon knows what you do online and offers up relevant content. Healthcare needs to start better utilizing these intervention tools in smart ways to help people lead healthier lives. If individuals tend to engage in risk behaviors in certain locations, we can use information from their phones to give them targeted health messages when they are close to that location.”
In terms of healthcare regulations and privacy, Dr. Kershaw admits that this can be tricky to navigate. But, he doesn’t see this as an insurmountable barrier to researching healthcare interventions: “If you are honest with people, build trust by taking the right steps to ensure privacy and let them know that their participation can help people, they are often open to opting into the research.”
Making It Work for Your Brand
Social behavior doesn’t just impact health—it impacts just about everything that humans do, including the adoption and use of brands, products and services. That’s one reason so many companies are focused on NPS. We assume, and research shows, that recommendations (a form of social behavior) are “catchy” and drive positive outcomes for our brands. The influence of social networks is also the impetus behind social media marketing programs executed on sites like Facebook and LinkedIn.
By understanding where and when category conversations happen and how information about your category flows through social networks, you can better address how, when and where to communicate with customers and prospects. And that understanding can help you identify targets and test and create programs to move your audience.
Look for my three-part series on using Big Data to sharpen market segmentation efforts in February. My goal is to help you get real value out of Big Data so please email me, if you’d like to recommend a topic for the series. Thanks!