Assessment of workingpostures by observation is a common practice in ergonomics. The present studyinvestigated whether monetary resources invested in a video-based posture observationstudy should preferably be spent in collecting many video recordings of thework and have them observed once by one observer, or in having multipleobservers rate postures repeatedly from fewer videos. The study addressed thisquestion from a practitioner’s perspective by focusing two plausible scenarios:documenting the mean exposure of one individual, and of a specific occupationalgroup. Using a data set of observed working postures among hairdressers, empiricalvalues of posture variability, observer variability, and costs for recordingand observing one video were entered into equations expressing the total costof data collection and the information (defined as 1/SD) provided by theresulting estimates of two variables: percentage time with the arm elevated<15 degrees and >90 degrees. Sixteen measurement strategies involving 1-4observers repeating their posture ratings 1-4 times were examined for budgetsup to €2000. For both posture variablesand in both the individual and group scenario, the most cost-efficient strategyat any specific budget was to engage 3-4 observers and/or having observer(s)rate postures multiple times each. Between 17% and 34% less information wasproduced when using the commonly practiced approach of having one observer ratea number of video recordings one time each. We therefore recommend observationalposture assessment to be based on video recordings of work, since this allowsfor multiple observations; and to allocate monetary resources to repeated observationsrather than many video recordings.