As Stack Overflow considers asking for any sort of recommendation to be off-topic, Stack Exchange Network launched Software Recommendations, a particular question-answering (Q&A) community for this purpose. Out of 182 such Q&A sites, Software Recommendations has been the least-answered site for the last 5 years, with only a 57% answered rate. We analyzed the complete data dump of Software Recommendations, containing data from its inception in 2014 to October 2025, to determine what makes it the least-answered site. We observed that incorrect tagging reduces the chances of questions being discovered by experts, and for this reason, certain topics receive better answers than others. Most questions are asked by users with a reputation of less than 300, while most of the answers are provided by those having a higher reputation. In total, 78.86% of users registered on the site act as silent observers and only 21.14% are involved in Q&A activities. Although knowledge-sharing activities on the site have decreased over this period, new users keep joining, which shows the need for software recommendations despite the increasing popularity of AI tools. Although similar questions are often asked and get closed as off-topic on sister sites, a very small proportion of such questions are migrated to Software Recommendations due to the lengthy migration process and reputation reversal on that site, which holds Software Recommendations back from attracting knowledgeable users and gaining popularity. Our findings suggest that the Stack Exchange community needs to revise its incentive mechanisms and devise ways to attract knowledgeable Stack Overflow users to the Software Recommendations site.
Building similarity graph...
Analyzing shared references across papers
Loading...
Arjumand Fatima
Onaiza Maqbool
Data
Quaid-i-Azam University
Building similarity graph...
Analyzing shared references across papers
Loading...
Fatima et al. (Mon,) studied this question.
www.synapsesocial.com/papers/69ba430d4e9516ffd37a3ec5 — DOI: https://doi.org/10.3390/data11030058