Modeling and predicting the growth and death of membership-based websites

Bruno Ribeiro

Abstract

Driven by the success of platforms like Facebook and LinkedIn, research often focuses on growth while ignoring the many ventures that fail. This study presents a mathematical framework using reaction-diffusion-decay processes to describe the Daily Active Users (DAU) of twenty-two websites. The model provides a mechanism to classify websites as self-sustaining or unsustainable and identifies whether growth is driven by marketing or word-of-mouth.

Introduction: Beyond Vanity Metrics

Traditional analyses of website popularity often suffer from survivor bias and a reliance on "vanity metrics" like total registered membership. Total membership only increases over time, masking the reality of user churn. This research argues that Daily Active Users (DAU) is the primary metric for assessing a site's health and social impact. By modeling DAU over time, we can capture the full lifecycle of a website, from its initial surge to its eventual stabilization or decline.

The Reaction-Diffusion-Decay Model

The proposed model treats users as particles in a chemical system, transitioning between three states:

  1. Active Members ($A$): Users currently visiting the site.
  2. Inactive Members ($I$): Registered users who are currently dormant.
  3. Non-members ($U$): The pool of potential future users.

The dynamics are governed by four fundamental processes:

  • Reactivation ($A + I \to 2A$): Active users encourage inactive ones to return.
  • Decay ($A \to I$): Users naturally lose interest over time.
  • Word-of-Mouth ($A + U \to 2A$): Active users recruit new members.
  • Marketing Diffusion ($U \to A$): External advertising brings in new users independently of the current active base.

Classification and Sustainability

A critical finding of the model is the Survival Ratio ($\beta/\alpha$), where $\beta$ represents the decay rate and $\alpha$ represents the reactivation rate. If $\beta/\alpha \ge 1$, the website is mathematically destined to decline, as the rate of losing active users exceeds the rate of bringing them back.

Classification of websites into self-sustaining and unsustainable categories based on reactivation (alpha) and decay (beta) rates.

As shown in the figure above, the model accurately classifies real-world platforms. Sites like Facebook and LinkedIn occupy the "self-sustaining" region, while others like Friendster and 12seconds.tv fall into the "unsustainable" zone.

Growth Signatures

The model also distinguishes between different "growth signatures" based on the initial driver of adoption.

Comparison of initial growth signatures: (a) media and marketing intensive vs. (b) word-of-mouth intensive.

Marketing-heavy growth ($\lambda \gg \gamma$) typically results in a convex curve as a large portion of the potential audience is reached quickly. In contrast, word-of-mouth growth ($\lambda \ll \gamma$) produces a classic exponential "S-curve," where growth accelerates as the active user base expands.

Key Results and Resilience

The model demonstrated high predictive accuracy, successfully forecasting Facebook's DAU stabilization point years in advance. Furthermore, it revealed that the underlying social dynamics of a site are remarkably resilient to temporary external shocks. For instance, political websites saw significant "DAU bumps" during the 2012 U.S. election, yet they quickly reverted to the trajectories predicted by their internal decay and reactivation parameters.

Conclusion

By shifting the focus from total membership to active engagement, this reaction-diffusion-decay model provides a robust tool for understanding the lifecycle of online communities. It highlights that in an attention-scarce economy, long-term survival depends less on initial marketing spend and more on the sustainable "reaction" of keeping users active.

Summary updated

Abstract

Driven by outstanding success stories of Internet startups such as Facebook and The Huffington Post, recent studies have thoroughly described their growth. These highly visible online success stories, however, overshadow an untold number of similar ventures that fail. The study of website popularity is ultimately incomplete without general mechanisms that can describe both successes and failures. In this work we present six years of the daily number of users (DAU) of twenty-two membership-based websites - encompassing online social networks, grassroots movements, online forums, and membership-only Internet stores - well balanced between successes and failures. We then propose a combination of reaction-diffusion-decay processes whose resulting equations seem not only to describe well the observed DAU time series but also provide means to roughly predict their evolution. This model allows an approximate automatic DAU-based classification of websites into self-sustainable v.s. unsustainable and whether the startup growth is mostly driven by marketing & media campaigns or word-of-mouth adoptions.
BibTeX
@inproceedings{ribeiro2013modeling,
  author = {Bruno Ribeiro},
  title = {{Modeling and predicting the growth and death of membership-based websites}},
  booktitle = {The Web Conference},
  year = {2013},
}
Figure