Using linked employer-employee data from the British Workplace Employment Relations Survey we examine how much of the variation in job quality is accounted for by establishment-level variation, and the relative importance of the establishment compared with occupation and employee characteristics. We do so for pay, six dimensions of non-pay job quality and overall job quality. We show that the establishment is the dominant explanatory factor for non-pay job quality, and as important as occupation in accounting for pay. Where you work accounts for between 38% and 76% of the explained variance in job quality, depending on the dimension. We also find that establishments which are 'good' on one dimension of non-pay job quality are 'good' on others. When we relate the estimated establishment effects (after allowing for the effects of occupation and of employee characteristics) to observed establishment characteristics, we find that non-pay job quality is greater in smaller establishments.