We consider differences in current job tenure of individuals using linked employee and workplace data. This enables us to distinguish between variation in tenure associated with the characteristics of individual employees and those of the workplace in which they work. The various individual characteristics are, as a group, found to be essentially uncorrelated with the workplace effect, however, this is not true for women and non-white employees. We find that the lower tenure associated with membership of these demographic groups is predominantly captured by workplace effects suggesting some degree of labour market segmentation in Britain.