The Data Deluge and Behavior Changeby David Kim

On February 27, 2010, the Economist dedicated an entire issue called Data, data everywhere to report on managing information in the 21st century. As the title suggests, the issue highlighted the deluge of data created, tracked, and stored on all aspects of human life. In fact, according to a 2008 study by International Data Corp, a market-research firm, around 1,200 exabytes of data will be generated this year. To put that number in perspective, this is more data collected than all of past history combined!

Given this deluge, the obvious question arises: how has the flood of data changed the way we behave? Unfortunately, answering this broad question is not a trivial pursuit and perhaps more appropriate for an in-depth sociological research paper. But a blog post may be more suited to discussing the mechanisms by which data and the underlying agents tend to affect behavior change.

One view of data focuses on is its agnostic nature. Data does not judge or discriminate. Whether it is qualitative or quantitative, it seeks to objectively measure and describe the world around us. Of course there is often human intent and subjectivity through which data is collected and recorded, but by-in-large, data merely encodes information about the world.

The type of behavior change resulting from simply reacting to this agnostic data presented is probably the most natural or “organic” form of behavior change. Think about the increasing amount of personal health data that is collected. Services such as 23andMe or Counsyl provide in-depth genetic information, making affordable a full profiling of someone’s underlying makeup. Seeing that her tests suggest a high chance of diabetes thanks to (or unthanks to) hereditary factors, a user may naturally be inclined to reduce sugar intake and avoid as much stress as our current high-paced society allows. Or someone using fitibit (a small device that tracks both the number of steps walked and biometrics about the quality of sleep) observing lagging physical activity on Wednesdays may choose to bike to work during those days rather than drive. Merely using the data before us as a way to inform decision making to change normal behavior represents a very “natural” response and is not necessarily unique to the data explosion we are experiencing in the digital era.
How about data that is actively being used to control behavior in a very purposed manner? This represents the diametric opposition to the aforementioned organic behavior change. I refer to this as behavior engineering. Whether it is marketers trying to increase loyalty to a product or advertisers trying to drive traffic to a website, behavior engineers use social data to produce a desired outcome.
Let’s talk about gamification, one of the most powerful tools of behavior engineering. In short, gamification refers to the use of game dynamics to influence behavior. Game mechanics such as points, levels, and leaderboards are used to stoke basic human desires such as achievement, self-expression, altruism etc. Branchout, a popular Facebook application for those seeking a more career/professional oriented social networking site (think Linked-in for Facebook) is a prime example of a company using game dynamics in order to influence users to spread the word about its services. Using progress bars and a ranking feature of various attributes of people in a user’s network, Branchout creates an addictive service where people willingly provide personal data to the application. And as Branchout collects more user data, it can tweak its gamification tactics to optimize traffic flow, for example, via AB tests.
Similarly, Foursquare, the location-based check-in service, displays several statistics meant to motivate people to certain actions. The “mayor badge” in particular is bestowed upon the person who most frequently “check-ins” to a certain destination. By providing data about the current highest number of check-ins, Foursquare engineers a highly competitive incentive system. It isn’t difficult to imagine a user who may choose to visit, for instance, a Coupa Cafe she isn’t really interested in aside from the additional check-ins that would move her closer to acquiring the coveted mayor badge at this particular coffee store (no offense intended to Coupa). In such a manner, gamification is used to engineer a deliberate social system where data is leveraged to perfect game design, one that seeks to actively and effectively influence user behavior.

We’ve swung from one side of the spectrum to the other: data as purely information without a controlling agent to a tool for behavior engineering. The middle ground, which I call behavior architecture (a term not coincidentally related to the notion of choice architecture in the book Nudge), involves a less deliberate form of behavior change. Rather than for direct influence, data is used to create an ecosystem for interaction. However, there is a sense of intentionality in that behavior architects must choose what data is collected and presented, which has implications for the way users behave.
Consider Amazon. The online marketplace boasts one of the best product recommendation engines in the world. Its secret? User data. Since its inception, Amazon has been a user-data driven company, collecting and recording purchasing decisions and promoting features such as user product reviews that help people make better decision for themselves. The gargantuan cache of data collected is ultimately leveraged to help customers discover what is best suited to them. So where’s the behavior change? Such an effective recommendation engine may influence the customer to purchasing more items than she originally intended or may prevent her from discovering completely new products by serendipity.
How about the most obvious social architect Facebook? With over 500 million active users, the company has one of the largest repositories of data on basically all aspects of social activity on the web. What does that translate to? It means that Facebook has constant feedback about the way it can design its services such that people can best connect with others. For instance, the default option for the Facebook feed displays top news from closest friends. In such a way, Facebook attempts to provide users with the most relevant updates about the most important friendships. This has implications for the ways people behave. The feed feature, for instance, may reinforce the strongest relationships within a network while preventing otherwise more natural development of friendships along tertiary edges. Essentially, Facebook can influence how these interactions develop via the choice of social data they choose to prioritize.
Whether it’s through natural means, more controlling behavior engineering, or a middle ground involving architecting behavior, a data-rich world has many implications for behavior change. Specifically, given the extent of influence behavior engineers and architects can have, it seems important if not necessary for these data gatekeepers to act in a way that is responsible. In particular, it will be interesting to see how the data revolution can encourage behavior change that is not only socially responsible but helpful, beneficial and ultimately for the better.