Abstract :
Irregular and unpredictable events are increasingly important design elements of several of the latest business intelligence technologies, such as complex event processing (CEP), business performance management (BPM), and the real-time enterprise (RTE). Theories of individual and organizational learning from irregular events - exceptions, interruptions, surprises, accidents, and so on - tend to conform to one of two contradictory patterns. I draw these theories together to understand the complementary processes by which learners derive knowledge and insight from irregular events. I identify contingency factors that bias learners toward one of the two cognitive modes - incorporation of multiple events into a generalized understanding, or expansion of individual events into rich analytical conversations - and propose "exception design" levers by which BPM dashboard implementers can adjust these factors and influence the way users create knowledge in business intelligence. A pilot experiment lends some support to the hypothesis that frequency and ambivalence are designable aspects of exceptions that affect learning. The author seeks feedback and suggestions for a more effective research design.
