The effect on muscle activity variation of combining different office tasks is not known. We recorded electromyography from the upper trapezius (UT), wrist extensor (WE) and lumbar erector spinae (LES) in 24 office workers during five productive tasks, and breaks. Minute-to-minute variance was then estimated, by simulations, in a reference ‘job’ consisting of 85% sitting computer work and 15% breaks, and in ‘jobs’ where sitting computer work was replaced by different proportions of the other office tasks and breaks. Replacing sitting computer work with sitting non-computer work increased estimated variance by, in median, 23% (UT), 19% (WE) and 0% (LES). Replacing it by other tasks, in particular standing computer work and non-desk work, was less effective in increasing variance. Thus, some combinations of office tasks have a slight potential to increase muscle activity variation in the shoulder and lower arm, but not the lumbar back, while others will be ineffective.