Google+ or Google-?: Examining the Popularity of the new OSN

Roberto Gonzalez, Rubén Cuevas, Reza Rejaie, Ángel CuevasarXiv 2012 · arXiv

TL;DR

This paper dissects the first year of Google+'s evolution, revealing that while it attracted a massive user base, the platform's growth was largely driven by implicit account creation rather than genuine user engagement. Despite its size, Google+ exhibits characteristics of a broadcasting platform, similar to Twitter, rather than a deeply interactive social network like Facebook, with a small fraction of users driving most of the activity.

What Problem Does This Paper Address?

The launch of Google+ in 2011 presented a significant challenge to the established dominance of Facebook and Twitter in the social media landscape. While Google+ reported rapid user growth, there was considerable debate about whether this translated into genuine user engagement and whether it could become a relevant player. This paper aims to address this by providing a comprehensive, measurement-based characterization of Google+'s structure, user activity, and connectivity evolution during its crucial first year, seeking to understand if users were actively joining and participating or simply being added to the platform.

Methods

The authors employed a large-scale measurement methodology to capture comprehensive snapshots of Google+'s Largest Connected Component (LCC) and a significant number of randomly selected users. Their approach involved a breadth-first search (BFS) crawling strategy, optimized by dividing the user ID space into zones for parallel crawling. To obtain random user samples, they leveraged the Google+ API's search function, carefully selecting popular and unpopular surnames to mitigate bias. User activity was captured by collecting all publicly visible posts and associated reactions (plusones, comments, reshares) from the LCC users. They also collected comparable datasets for Twitter and Facebook to enable cross-platform comparisons.

Datasets

  • Google+ LCC Snapshots: Multiple snapshots of the Largest Connected Component of Google+ were collected over a year (e.g., LCC-Dec, LCC-Apr, LCC-Aug, LCC-Sep, LCC-Oct, LCC-Nov).
  • Google+ Random Datasets: Random samples of Google+ users were collected using API searches based on surnames (Rand-Apr, Rand-Oct, Rand-Nov).
  • Google+ Activity Data: Public posts and reactions from LCC users over a 15-month period (July 2011 - September 2012).
  • Twitter Datasets: Profile (TW-Pro), connectivity (TW-Con), and activity (TW-Act) data for Twitter users.
  • Facebook Datasets: Profile (FB-Pro), connectivity (FB-Con), and activity (FB-Act) data for Facebook users.

Key Insights and Ideas

  • Implicit Account Creation Drives Growth: The dramatic increase in Google+'s user base, particularly the large number of "singletons" (users not connected to the main network), suggests that many accounts are created implicitly through integration with other Google services (like Gmail), rather than through active user choice.
  • LCC Shrinks Relative to Network Size: While the absolute size of Google+'s Largest Connected Component (LCC) grew significantly, its relative size decreased over time, indicating that the majority of users were not actively connecting within the core network.
  • Broadcasting, Not Deep Interaction: Google+ exhibits characteristics of a broadcasting platform, similar to Twitter, where users primarily share information. This is evidenced by the low fraction of bidirectional relationships and the similarity in connectivity features to Twitter, contrasting with the more interactive nature of Facebook.
  • Activity Skewness is Extreme: A very small fraction of Google+ users are responsible for the vast majority of posts and, even more so, for receiving reactions. This "long tail" phenomenon is more pronounced for reactions, suggesting a highly concentrated engagement pattern.
  • Maturity in Connectivity: Despite continued user growth, key connectivity features of the Google+ LCC, such as degree distribution, have stabilized, indicating the network has reached a level of maturity in its structural properties.

Key Results

  • The LCC of Google+ grew at an impressive rate of 150-200K users per day.
  • However, the relative size of the LCC dropped from 43% in April 2012 to 32% by November 2012, with singletons making up 66% of the network.
  • Less than 9% of Google+ users generated at least one public post, and less than 25% provided any public attributes.
  • The top 10% of users contributed 80% of all posts, while the top 1% of users received roughly 80% of comments and 90% of plusones and reshares.
  • Google+ users were found to be significantly less active than users on Facebook and Twitter, with a typical activity rate of 0.08 posts/day compared to 0.19 for Facebook.
  • The fraction of bidirectional relationships in the LCC stabilized around 21.3%, similar to Twitter's 22%, but significantly different from Facebook's fully bidirectional nature.

Limitations and Future Work

The authors acknowledge that their study focuses on publicly visible activity, and private posts are not captured due to technical and ethical limitations. They also note that while their random sampling strategy is robust, there's a slight possibility of miscategorizing users as singletons if their friends/followers also have private lists. Future work could explore the reasons behind user departure from the LCC, investigate the impact of Google+'s "Circles" feature on user engagement, and delve deeper into the dynamics of user-generated content and its consumption. The paper suggests that while Google+ can attract users, fostering meaningful engagement remains a significant challenge in the competitive social media landscape.

Summary updated

BibTeX
@misc{gonzalez2012google,
  author = {Roberto Gonzalez and Rubén Cuevas and Reza Rejaie and Ángel Cuevas},
  title = {{Google+ or Google-?: Examining the Popularity of the new OSN}},
  year = {2012},
  eprint = {1205.5662},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1205.5662},
}
Figure