An empirical analysis of human performance and error in process model development

Process models capture important corporate know-how for an effective Business Process Management. Inconsistencies between process models and corporate reality are a common phenomenon in corporate practice. Human performance in process model development is a major source for these inconsistencies. In this contribution, a human performance analysis of process model development paying special attention to the concept of human error was conducted. It was found that the frequencies of the omissions and erroneous executions of notation elements are significantly higher for novices than for experienced modelers. Moreover, experienced modelers inherently adhere to a verb-object activity labeling style. The overall empirical results indicate that experienced modelers achieve higher process model quality with less expenditure of time than novices.