The Sustainability of Corporate Wikis: A Time-Series Analysis of Activity Patterns

Abstract

While existing theoretical frameworks describe collective technology adoption patterns, they provide little insight regarding the expected patterns of wiki activity within projects. Another impediment to the study of wiki sustainability is the absence of time-series analysis methods that are suitable for the unique patterns of wiki activity logs. The primary goals of this study are to: (i) develop a novel method for analyzing wiki edit activity logs, (ii) reveal the temporal patterns of corporate wiki edit activity, and (iii) study the factors impacting wikis' sustainability. A validation of our proposed method demonstrates that it is superior to the baseline algorithm in the face of noisy data. Our empirical study combines wiki system edit activity logs with a survey of users' perceptions, and explores 33277 distinct wiki applications within one global organization over the first 5 years of wiki operation. Our results reveal six different prototypical wiki activity patterns, and show that most corporate wikis become inactive after a relatively short period. Findings from the user survey show that users of sustainable wikis are more satisfied with the wiki system and its contents, and feel that the wiki provides them with a sense of community and productivity enhancements

Keywords: Time Series, Corporate, Wiki, Life cycle, Sustainability, Edit activity patterns, Clustering