Quantitative characterization of postural load on the back should describe exposure patterns among workers and factors affecting these exposure patterns. This article presents general guidelines for designing appropriate measurement strategies; how to obtain detailed data with an applicable measurement method, what sampling strategy should be applied, which frequency and duration of measurements are required, and how differences between workers, shifts and tasks are addressed. Formulae are available to evaluate the best trade-off between the reliability of a measurement device and repeated measurements and between the number of workers to be monitored and the number of measurements per worker. More difficult questions relate to collective assessments versus individual assessments and duration versus frequency of measurements. Methods are described for statistical modelling of data on trunk flexion and rotation. A two-way analysis of variance was applied to assess the principal sources of variation in back load among workers in the woodworking industry. A bootstrapping technique was used to evaluate the minimum number of measurements required to arrive at an unbiased estimate of the average exposure to trunk flexion in two occupational groups. It is advisable to conduct a small pilot study to determine the essential features of the measurement strategy before undertaking a large study comprising many workers and many work situations.