Recency or Quality: Final Interactions Overpower Expectations in Subjective Usability Ratings

Subjective usability scales, such as the SUS and SEQ, are assumed to provide global assessments of user experience. However, emerging evidence suggests that the final interaction has a disproportionate impact on these subjective ratings, potentially leading to inaccurate global usability scores. This study investigated whether establishing a stable expectation of quality through multiple interactions could buffer against such recency effects. Participants completed a series of search tasks in which the quality of the initial and final interactions was systematically varied. Results revealed that the quality of the final interactions dominated the evaluation, regardless of the user’s initial experiences with the website. For instance, even after several positive interactions, final negative interactions reduced usability ratings to levels indistinguishable from those of an entirely bad design. These findings reveal a strong biasing effect that could inadvertently trigger critical redesign decisions based on transient end-of-session frustrations rather than on objective system quality.