The Virtual Behavior Laboratory

-- By TheodoreSmith - 09 Jan 2009

Table of Contents

Introduction

Over the last 15 years, the World Wide Web (WWW) has developed into a popular means to communicate, share information, and engage in commerce. With the Web’s increasing ubiquity has come ever more sophisticated means for webpages and ISPs to collect and collate usage data. Even as these data effect a revolution in the specificity of targeted marketing techniques, a possible second transformation threatens: the increasing popularity of Massively Multiplayer Online Environments (MMOs), typified in games such as World of Warcraft. Although this technology is still developing, the immersive nature of the experience, coupled with the ability of a server to capture and record every aspect of the game, provides marketers with a sophisticated online laboratory for the development of targeted behavioral models and marketing.

The Web

At a high level of abstraction, the structure of the WWW protocol is built around a framework of requests and responses: an end user sends a request to the web server and the server responds with the desired information. Although this is a simple and efficient structure for information retrieval, the open nature of the Internet permits these data transactions to be recorded by any entity placed along the path of communications. Once collected, these data may be associated back to the original user and collated to build a model of preferences and interests. Whether this data is eventually used as a marketing tool by the gathering entity or simply sold outright, its existence represents both a powerful tool for targeted marketers, and a danger to the autonomy of the WWW user.

While the ability to collect WWW usage data presents a powerful tool for the marketer, the Web is in many ways not an ideal source of such data: a content provider can generally only track the user by the requests his browser makes to the website. Although discrete data can be extracted from this arrangement – what the user is interested in, what advertisements the user fancies, etc. – compiling this data into a useful behavioral model is more difficult, and requires the interested party to fill in the gaps between clicks. This should not be considered an intractable weakness of the medium: it may very well be possible to capture detailed behavioral data through a carefully planned and monitored site. In general, however, current Web marketing tends to focus simply on preference data: matching advertisements with a user’s interests.

MMOs

Modern MMOs are virtual environments in which a number of users concurrently interact with a visual representation of a world through a digital avatar. These worlds differ from the World Wide Web in a few important respects: first, the world of the MMO is intended to be immersive, allowing the user’s avatar to directly interact with the virtual world. Although any virtual environment must at some level be an abstraction, the purpose is to give the user a sense of existence within the world or game. This immersive quality creates an excellent playground for the study of human interaction, not only placing the user in a situation where they are reacting authentically to information presented by the environment, but also giving the developer complete control over the user’s experience.

Second, insofar as the MMO exists entirely as a digital abstraction, it is possible to monitor all aspects of a user’s progress in the world. While some information may be gleaned from web page usage, many aspects of the user’s experience may be outside the reach of the web site developer; pages requested are not always pages viewed, and the actual experience of the web, particularly with the advent of tabbed-browsing, can be very non-linear. In a virtual environment, every aspect of the user’s experience may be designed, controlled, and recorded in linear sequence.

These two aspects of the MMO – full control over a user’s environment, and the ability to track all details of a user’s interaction – provide a potential marketer with a ready laboratory for the testing and development of models and marketing approaches. The immersive environment of the MMO has the potential to provide a marketer with a vast source of specific behavioral data: how a user tends to make decisions, the information sought before a purchase, the influence of other people or authorities, and the desirability of rewards.

Possible Futures

Current MMOs are typically financed by monthly fees, and are not publicly in the business of collecting or selling behavioral data on their subscribers; however, as the user base familiar with these environments grows, the concept of the multiplayer online environment is likely to be expanded into other areas and business models. As these online environments become more popular and seek new streams of revenue, behavioral data is more and more likely to become a commodity. Attempts at bringing the real world into the MMO medium are still atypical; however, the convergence of web style commerce and the MMO presents the dilemma in its clearest form: a virtual retail environment capable of recording every word spoken to or by a particular user, capturing how he pauses at each item and advertisement, and tailoring the details of the store to precisely match his shopping style, tastes, and weaknesses, as compiled from data gleaned from hundreds of similar interaction between the user and other vendors.

Conclusion

Although the scenario described above is a plausible projection based on current trends, the internet is littered with similar incorrect predictions; the MMO’s eventual mainstream appeal and the precise trajectory of the genre’s development remain a mystery. Whether or not the MMO itself plays a significant role in the future evolution of the internet, the medium highlights a pair of important trends: the ever decreasing granularity and increasing specificity of data collection, and the ease of building and testing targeted behavioral models through increasingly immersive online interfaces. The issues raised by the collection and sale of internet usage data are not going away, and are likely to grow in subtlety and degree as our methods of communications become ever more immediate and ubiquitous.

Interesting paper, Ted. I think the potential scenario you articulate is plausible; there's a clear impetus in the private sector to figure out how to better collect and apply behavioral data. Even if virtual words aren't the particular mechanism that comes into favor, knowing how a potential customer thinks and the most direct way to influence him is of tremendous value to a lot of companies, so you can bet a lot of thought and effort will go into designing better online laboratories.

What I'm personally struggling with is articulating the harm that comes from commercial exploitation of all this knowledge. Is the individual's autonomy threatened when corporations get very good at selling him things? Or is the harm to society as a whole--rampant capitalism, bubbles and recessions, that sort of thing? Isolating the undesirable outcomes that we're concerned about would help a lot when thinking about issues of privacy.

-- AndreiVoinigescu - 09 Jan 2009