We used a combination of inexpensive camera equipment and statistical analysis of optical flow patterns to analyse the behaviour of 24 commercial broiler (meat) chicken flocks, Gallus gallus. Individual birds were not tracked or marked but the skew and kurtosis of flow patterns produced by the collective movements of the flocks were significantly correlated with key welfare measures such as % mortality, numbers of birds with hockburn (damaged leg skin) and abnormal walking behaviour (poor gaits) in individual birds. These correlations were already apparent in birds as young as 15–20days. Optical flow patterns provided an information-rich link between flock and individual that could be important in the development of new ways of assessing the welfare of, and managing, broiler chickens. It could also have wider application to the study of other animal groups as an alternative to more invasive or intrusive methods. ► Optical flow provided a noninvasive way of monitoring welfare in broiler chickens. ► Uniform optical flow indicated lower mortality, healthier legs and better walking. ► More extreme flow predicted higher mortality, blemished legs and lameness. ► Flow patterns predicted welfare days or weeks in advance. ► Optical flow could improve broiler welfare through better management.