Sailer combined a staff survey at the UK headquarters of a global technology company with an isovist analysis of every desk, which counts how many other desks fall inside a person’s field of view. The analysis runs on 167 desks with a matched survey response; five more were excluded because the seat type was missing. The average desk saw 66 others, and the most exposed saw 223. Adding half a standard deviation to that count lowered the odds of a favourable rating for working productively by 58 per cent and for concentrating by 45. Sharing information fell 35 per cent, planned meetings 33 and team identity and cohesion 28. Every one of the collaboration measures moved the same way as the concentration measures. The case for taking out the partitions rests on them moving in opposite directions. People seated next to a wall rather than next to a window were less satisfied with their ability to work productively by a factor of 4.3.
At one corporate headquarters, 52 employees wore badges that recorded whom they faced. Ethan Bernstein and Stephen Turban counted 5,266 minutes of face-to-face interaction over 15 days before the move to open plan and 1,492 after, a fall of 72 per cent. At a second, 100 employees in 1,830 pairs spent 67 to 71 per cent less time face to face. Instant messages rose 67 per cent and the words in them 75 per cent at the first firm, and email between pairs rose 22 to 50 per cent at the second.
Mahbub Rashid separated the two things a layout does, using space syntax and observed behaviour in four moderately large offices. Spatial layouts had consistent effects on how people move and inconsistent effects on visible copresence and on face-to-face interaction. Visible copresence predicts whether two people talk; movement through the building does not. The abstract available here gives the four offices and not how many observations were taken in them.
Rianne Appel-Meulenbroek asked 138 employees of one research organisation to keep a week-long diary, which recorded 918 unplanned knowledge-sharing meetings. Analysed against the building layout, inter-visibility and proximity were most strongly associated with where knowledge got shared. Sharing a room and being able to overhear were associated with how it got shared, meaning what was discussed and whether it was intentional. The differences appeared at a very local level inside the building rather than between wings or floors.
Leroy Gonsalves spent 14 months inside a large team-based sales company through an office redesign, watching what happened when assigned cubicles gave way to an unassigned mix of spaces. The old arrangement made workers permanently accessible. The new one changed what colleagues could know about where someone was and whether they wanted to be found, which gave people ways to hide and ways to signal availability by choosing where to sit. The result was more asynchronous communication and less unwanted interruption. The workers described the change as control over their own availability. Bernstein’s badges recorded the same withdrawal from outside, as a 72 per cent fall in face-to-face time. The abstract available here reports the fourteen months and not how many people were observed.
Every one of these studies is a single organisation. Bernstein and Turban put badges on 152 people and Sailer an isovist on 167 desks. Appel-Meulenbroek’s 138 employees logged 918 meetings themselves. Rashid observed four offices without giving a count. Gonsalves spent 14 months inside one company.