Can the growth and decline of collective attention to social media services be predicted?
Understanding the dynamics of collective human attention has been described as a key scientific challenge of the information age.
A recent paper1 explores an aspect of this problem by analysing to what extent the general dynamics of collective attention apparent from search frequency data for searches for social media services can be modeled mathematically.
Google Trends data for 175 different social media services was used, and for each service data from 45 countries as well as global averages was collected. The more than 8,000 time series obtained were analysed using diffusion models from the economic sciences.
Diffusion models, which are well established in economics, were used in the study because of their conceptual simplicity, being designed to characterise time series in terms of everyday concepts such as the propensity for collective attention to grow and to decline.
Three diffusion models were fitted to the data and their performance evaluated: the Bass, shifted Gompertz, and the Weibull.
The most important results of the analysis are:
- Economic diffusion models provide accurate and statistically significant explanations of general trends in aggregated search frequency data which summarize how collective attention to social media evolves over time.
- Collective attention to social media evolves according to simple and highly regular dynamics of growth and decline.
- Collective attention to social media evolves globally similarly and independent of regions of origin or cultural backgrounds of crowds of Web users.
- Most social media services are able to attract growing collective attention for a period of 4 to 6 years before user interest inevitably begins to subside.
The study authors further conclude that:
…it appears that public attention to social media hinges on perceived novelty. In other words, the more a crowd of users gets used to a service or the less novel it appears, the faster it looses its appeal. These are the characteristics of hype cycles. The temporal behavior exposed in our analysis is therefore well in line with everyday experience and aptly summarized by the statement that what goes up, must come down.
Twitter is the latest social media service to illustrate this reality, having been overtaken by messaging services which are only at just at the beginning of the growth phase of their growth and decline cycle.
As well as being of interest to marketing and public relations professionals, the study findings have relevance to knowledge managers. The eventual decline of the social media services we use for knowledge exchange and transfer needs to be anticipated and planned for in advance.
This arguably should have been done in the case of Yammer. The current decline of Yammer could have been anticipated, and new services established to take its place before it reached its peak.
Image source: Social Media apps by Jason Howie is licenced by CC BY 2.0.
- Bauckhage, C. and Kersting, K. (2014). Strong Regularities in Growth and Decline of Popularity of Social Media Services, arXiv:1406.6529 (cs.SI). ↩
Also published on Medium.