Purpose -This paper explores which factors may explain the high or low percentages of satisfied employees in offices with shared activity-based workplaces. Design/methodology/approach -The paper compares data on employee satisfaction from two cases with remarkably high satisfaction scores and two cases with significantly lower satisfaction scores (total N = 930), all of the same organisation. These cases were selected from a database with employee responses to a standardised questionnaire in 52 flexible work environments. In the four case studies, also group interviews were conducted. Findings -Overall, there are large differences in employee satisfaction between cases with, at first sight, a similar activity-based office concept. The main differences between the best and worst cases regard employee satisfaction with the interior design, level of openness, subdivision of space, number and diversity of work places, and accessibility of the building. Employee satisfaction shows to be influenced by many physical characteristics of the work environment and by the implementation process. Satisfaction with the organisation may have an impact as well.Research limitations -Almost all cases regard Dutch organisations. Due to the lack of quantitative scales to define the physical characteristics of the work environment, the study is mainly descriptive and explorative and does not include advanced multivariate statistical analyses. Practical implications -The data revealed clear critical success factors including a supportive spatial lay-out to facilitate communication and concentration, attractive architectural design, ergonomic furniture, appropriate storage facilities, and coping with psychological and physical needs such as privacy, thermal comfort, daylight and view. Critical process factors are the commitment of managers, a balance between a top-down and a bottom-up approach, and clear instructions on how to use activity-based workplaces. Originality/value -The study connects descriptive research with inductive reasoning to explore why employees may be satisfied or dissatisfied with flex offices. It is based on a combination of quantitative survey data from 52 cases and a closer look at two best cases and two worst cases based on qualitative data from interviews and personal observations. The study has high practical value due to the integral approach that incorporates many items of the physical environment and context factors like the implementation process.