Task-Based Open Card Sorting: Towards a New Method to Produce Usable Information Architectures.

Open card sorting is the most widely used HCI technique for designing user-centered Information Architectures (IAs). The method has a straightforward data collection process, but data analysis can be challenging. Open card sorting has been also criticized as an inherently content-centric technique that may lead to unusable IAs when users are attempting tasks. This paper proposes a new variant of open card sorting, the Task-Based Open Card Sorting (TB-OCS), which considers users’ tasks and simplifies data analysis. The proposed method involves two phases. First, small groups of participants perform classic open card sorting. Then, each participant performs findability tasks using each IA produced by the rest participants of the same group and their first-click success is measured. Analysis of the collected data involves simply calculating the first-click success rate per participants’ IA and selecting the one with the highest value. We have also developed a web-based software tool to facilitate the conduction of TB-OCS. A within-subjects user testing study found that open card sorting produced IAs that had significantly higher first-click success rates and perceived usability ratings compared to the IAs produced by TB-OCS. However, this may be due to parameters of the new method that require finetuning, thus further research is required.