Drifting away of actions from prescribed procedures

This research examines a particular kind of routine change where a decoupling of actions from the prescribed procedures is consciously and mindfully introduced to benefit the organisation and to make work easier, locally efficient, and more effective. I draw on Snook’s (2002) conceptualisation of practical drift to define my phenomenon. When procedural demands do not fit the practical or situational demands, people adjust or alter the recommended procedures, routines, and workflows to better fit the local needs. This research uses a case-based inquiry to examine 197 cases of drift happening within large multinational organisations from the Oil and Gas, Manufacturing, and Aerospace sectors to theories of why ‘Practical drift’ happens and what the impact of this on safety. Using a mixed-method data collection technique of Repertory Grids (RepGrid), 31 middle and senior managers were interviewed. It resulted in 262 RepGrid ‘Constructs’ related to the events of drift. The data was firstly analysed qualitatively using a bootstrapping generic content analysis technique that pools construct into meaningful higher-order categories. Two quantitative analyses followed this: a RepGrid-specific statistical analysis, called the Average Normalised Variance (ANV), to identify key constructs and a Boolean minimisation-based Qualitative Comparative Analysis (QCA) to identify factor combinations associated with safety negative outcomes. This research finds that drifted actions happen because actors have an operational objective or ‘Purpose’ that triggers drift. Secondly, there exist some contextual conditions that facilitate replacing a recommended procedure with an alternate action. Safety findings reveal that some factors like ‘Risk awareness’ and ‘Local decision flexibility’ are more likely to deliver safety negative outcomes. In contrast, the factor ‘save time’ was found to be correlated to safety positive outcomes. Building on these findings, I propose three theoretical models. First is a factor model that identifies a set of factors causing drift. These factors are ordered by their relative influence, extending our understanding of the purpose and contextual conditions associated with drift. The second and third models link factors from the first model to different safety outcomes. These models add to the extant literature on practical drift and routine change by identifying (i) factors that link to safety positive, neutral and negative outcomes and (ii) the combination of key factors having more potency to deliver safety negative outcomes. The relevance of these findings for practice is that the research raises the issue of drift being an operational reality and motivates the organisations to address the drift causing factors. Furthermore, this study paves the pathway for future studies to establish causal relationships among the configuration of constructs discovered in this research.