Human error analysis in software services industry

Abstract

Quite a number of disasters have had the root cause as human error. While the erroneous act is unintentional, the consequences have been multifold causing loss of lives and money. Over the years the human error has gained significant importance and several studies have been performed to understand the cause. We usually see such devastating incidents in industries like Aviation, Nuclear and Medicine. The underlying cause may be due to human-machine interaction, human computer interaction, erroneous software, cognitive reasons etc. While automation appears to be a catchall remedy for human errors, one may not be able to automate in some cases. Human interaction may be minimized but complete elimination may not be applicable in all cases. It is therefore necessary to focus on the design to prevent potential human errors in the system. There are broadly two classifications of error types Human and Technical/Process. There are models and methodologies to cater to these types of errors. The Technical/Process error is usually driven by business models, technology. Human error analysis is based on cognition framework and there are a lot of studies performed in industries such as aviation, nuclear science. However, detailed analysis of human error in software services has been uncommon. This study is positioned in this place to look at human error analysis methodologies and framework that can be adapted from other industries into Software Services. Several studies have been performed especially in the Aviation industry to dissect the cause of accidents and many models and frameworks have emerged. Some of the noted ones mostly related to the field of Human Reliability Assessment (HRA) are THEA (Technique for Human Error Assessment), CHLOE (Collaborative Human Learning Object Exchange), Human Error Prediction methods such as HEART(Human error assessment and reduction technique), THERP (Technique for Human Error Rate prediction) etc. These methods provide a structured approach to analyse the human error and provide insights on avoiding the same at design and strategy stages. In a Software Services organization the customer is serviced using a combination of one more software applications and tools. There will be human-computer interaction. The users of these applications are usually from the software industry providing services to customers through that software. Typically such services are provided through service requests or tickets logged by customers for issues faced by them. The resolution to such requests would require a series of steps to be planned, performed and validated. There are chances that the software analyst performing such activities could miss a few steps or perform the step incorrectly. This may lead to loss of business for the organization and the customer, reduced customer confidence and revenue loss. The current study is an attempt to adapt proven methodologies in Human Reliability Assessments area into such scenarios in software services and to come up with models that applies to this industry. Due to time constraints these models could not be implemented as a part of this study. However, as a method of validation of its usage and applicability to software services, experts were interviewed and this study was presented and they have found this to be workable.