Improve Web System Navigation with Social Network Analysis (SNA)

Posted February 28, 2017 by Damion Mullins

Today, we use the term social network mostly to refer to applications and services that allow people to connect, such as Facebook or Twitter. The origin of the term, however, was coined in social and behavioral sciences to describe the relationships among individuals and groups. Since then, a methodology called Social Network Analysis (SNA) has emerged as an invaluable statistical tool with immediate benefits in many fields of study, including the commercial IT industry.

Social Network Analysis (SNA) is the mapping and statistical measuring of relationships and flows between various entities such as people, groups, and organizations. Using SNA, developers can investigate these relationships and flows through the concepts of nodes and links. In an SNA visualization, for example, nodes could be used to represent people, and links could be used to represent the connections between those people. The image below shows an example of a social network represented as a system of nodes and links.

One of the most relevant uses of SNA in the IT industry is in recommender systems, such as those used by companies like Netflix and Amazon. Netflix’s user interface shows recommended movies based on a variety of related viewer characteristics. Similarly, Amazon’s user interface provides product recommendations based on characteristics of products and their buyers. Key to these recommendations are the relationships that exist between products, as well as the relationships that exist between users. SNA is the primary tool to leverage these relationships.

In an SNA visualization, the volume of connections between nodes is indicated by the thickness of the link, whereas the importance of a node relative to other nodes is indicated by its diameter. For example, if Product A is connected to Product B via a relatively thick link, this indicates that there is a significant relationship between these two products. This also means that when a user is viewing Product A, the recommender system will likely display Product B, and vice versa. This type of SNA measurement is known as degree centrality, or the degree of adherence to the mean.

SNA can provide value to any company, no matter the size. In the 2005 case study “Discovering Social Networks from Event Logs” (Aalst, Reijers, and Song), SNA was used to show statistically significant relationships between end users in a business system. After discussing the results of the study, the end users reported that the data provided useful and novel information about the duties and roles of user employees and their reliance on the system.

So how can SNA improve web system navigation? Try to think of menu options on a web application or system as nodes, with each click creating a link between each menu and sub-menu option. Relationships between the web pages would be represented as nodes of different sizes and as links of different thickness, presence, and direction. For example, the “Home” button menu option would be represented by a very large node with many links coming from other nodes, because end users typically navigate to the home screen multiple times when using a web application. Likewise, a “My Profile” menu option that provides personal account information might also have a significant number of links coming from the home page.

A further benefit is realized whenever SNA identifies menu options with significant click behavior that were not previously thought meaningful. For example, take a university website that allows computer science students to register for a front-end web development course. The university may offer elective courses outside of this major that may be of interest to computer science students. SNA can identify the current trends among like-minded students, identify the significant relationships between the two courses, and allow the university to immediately recommend a high-interest elective based on these analytics.

As in the example above, SNA can be applied to improve web system navigation by adapting the system’s user interface to render the identified menu options as “recommended” or “frequently used” for quick navigation, similarly to how Amazon recommends products. A menu option recommender system powered by SNA could help new end users navigate through large web applications or systems more easily, as recommended options would be automatically displayed rather than buried in convoluted static menus.

An SNA-powered menu option recommender system is just one of the many ways SNA can be integrated into web applications and systems. SNA results are predictive, statistically significant, and dynamic, resulting in information that is personally tailored to end users.