No abstract

This paper uses data on invoice processing in four organizations to distinguish empirically between two competing theories of organizational routines. One theory predicts that routines should generate patterns of action that are few in number and stable over time, and that atypical patterns of action are driven primarily by exceptional inputs. The competing theory predicts the opposite. By modeling the routines as networks of action and using a first-order Markov model to test for stationarity, we find support for the competing theory. The routines generated hundreds of unique patterns that changed significantly during a five-month period without any apparent external intervention. Changes did not appear to reflect improved performance or learning. Furthermore, we found that exogenous factors (such as large invoices from unusual vendors) are not associated with atypical patterns of action, but endogenous factors (such as the experience of the participants) are. We also found that increased automation can increase variation under some circumstances. These findings offer empirical support for endogenous change in organizational routines and underscore the importance of the sociomaterial context in understanding stability and change.

This paper demonstrates that patterns of action are a fruitful basis for an empirical science of organizational routines by analyzing data from the invoice processing routines in four Norwegian organizations. Invoice processing is a highly institutionalized activity, governed by accounting rules and subject to audit. These four organizations use the same technology for the same task, yet the patterns of action generated by the routines are significantly different. We analyze the patterns in terms of individual events, pairs of events, and whole sequences of events, and at two different levels of abstraction. We discuss the implications of this approach for theories of routines as genes, and for an empirical science of organizational routines.

We reexamine the assumptions of current theory to update and extend the concept of task complexity to tasks that include multiple actors at any level of analysis. Tasks can be modeled as networks of required actions and information cues carried out or processed by particular actors. Counting pathways in the task network provides an index of task complexity that incorporates insights from organization research but is more consistent with contemporary complexity science than prior approaches and better reflects the exponential nature of the phenomenon. The revised concept of task complexity can be readily used as an independent or dependent variable, and it can be used to compare observed and idealized descriptions of tasks. We discuss its implications for developing theory.

In recent years, organizational routines have been studied in a wide variety of settings, including law, medicine, accounting, and engineering. This fieldwork has led to a broader understanding of organizational routines as repetitive, recognizable patterns of interdependent action, carried out by multiple actors. Routines are seen as practices that are situated in a social/material context. Within an organizational routine, individual actions are situated in a broader pattern of actions that can be represented as a network. Recognizing patterns of interdependent action as a unit of analysis entails a research paradigm that has implications for a range of topics in organizational behavior.

Research on expertise has shown that nonexperts may sometimes outperform experts. Some researchers have suggested that superior performance by experts depends on the match between the experts' cognition and the demands of the task. The authors explored this issue using a quasi-experiment set in an organization. They examined how 3 sets of similar tasks that differ in their type of complexity can lead to differences in task perceptions and performance among experts, intermediates, and novices. The results suggest that experts and novices pay attention to different aspects of a task and that this affects both their perceptions of task complexity (i.e., task analyzability and variability) and their performance on the task.

No abstract

Similar to practices in top management positions worldwide, there has been an increasing tendency in recent decades to fire football managers when the team does not perform to the stakeholders' expectations. Previous research has suggested that improvements after change of manager are a statistical artefact. Based on 12 years of data from the Norwegian Premier League, we conduct a natural experiment showing what would have taken place if the manager had not been fired. In this case, the performance might have improved just as well and even quicker. Building on theories in expertise and decision making, we explore the data and argue that decision makers may be fooled by randomness and learn wrong lessons about team leadership. Our analyses support a post-heroic view of team leadership as an emergent, output variable. Exaggerated focus on the individual manager may ruin long-term performance. Practical implications are discussed.