Purpose: This article addresses the question of why organizations do not uniformly apply pay for performance (PFP) throughout the organization, focusing on the wider occupational structure in which they and the jobs they create are embedded. We propose a model of ‘occupational differentiation’ whereby the probability of a job within a given organization having PFP increases with levels of monitoring difficulty and requisite human asset specificity characterizing the occupation to which a job belongs, being highest in occupations characterized by high levels of both (generally managerial and professional occupations). Design/methodology/approach: Using the Workplace Employment Relations Survey (a nationally-representative matched employer-employee dataset for Britain), this paper investigates this question for all 350 occupations delineated by the UK’s Office for National Statistics using regression methods that adjust for other confounding factors such as demographic factors and workplace fixed-effects. Findings: We find organizations ‘occupationally differentiate’ the use of PFP in ways consistent with the model i.e., PFP is most likely to be found in occupations characterized by both high monitoring difficulty and high requisite human asset specificity (mainly managerial and professional occupations) and least likely in occupations scoring low in both. The finding holds across PFP types (individual, group, organizational), whether organizations are large or small, and hold across most industrial sectors. Research limitations/implications: The main implication of this study is that organizations appear to be taking into consideration whether the wider profession to which a job belongs when implementing PFP, irrespective of their own HRM strategies and organizational context. There are a few limitations to this study, with the main one being that empirical support is only found in the private sector. The public sector appears to be beyond the reach of the model, where PFP is generally rarer. A second limitation is that the dataset is from 2011 and only covers a single country. Practical implications: Given organizations appear to be implementing PFP based on occupation, this may lead to equity concerns as different groups are being treated differently within organizations based upon their occupational group. Social implications: Since PFP jobs tend to pay more than non-PFP jobs and PFP prevalence has been growing, by being more likely to implement it for generally high-paid groups (generally higher managerial and professional occupations), PFP may contribute to wider pay differentials within and between organizations. Originality/value: By introducing the occupational-level of analysis and the differential nature of tasks across occupational groups, the model offers a new midrange, sociological perspective to understanding intra-organizational dynamics in PFP-use, and potentially HR practices more broadly.