Knowledge transfer model of team capability in non-collocated software architecture development

Software architecture is all about making design decisions based on the requirements specified by the users. Knowledge transfer (KT) is crucially essential as for making these design decisions, many factors and inputs need to be carefully considered and accounted. However, not much is known about KT in software architecture development, a setting that is very much knowledge intensive. Unclear flow of KT and insufficient explanation on how KT may influence teams’capabilities are identified as the problems seeking for immediate delineation. As awareness of the importance of KT in software development has increased, together with a growing interest in related capabilities, this research proposes to investigate how the KT unfolds between analyst and software architect teams in non-collocated software architecture development. Our interest also lies into the KT factors that might positively influence teams’ capability in terms of team and task familiarity. 150 participants consisting of analysts, software architects and a few project managers have been interviewed in two distinct phases of semi-structured interview sessions for data collection purpose. We describe and characterize KT through five key factors or constructs into a model: 1) the areas of knowledge used and exchanged, 2) the interdependency between teams, 3) the utilization of knowledge used and exchanged, 4) the mediums used for KT, and finally 5) the external conditions surrounding KT. The findings reveal confirmatory with our model. We used linear regression analysis for hypotheses testing, to check whether these five factors influence teams’ capability in terms of team and task familiarity, as well as their influence direction. The results have proven the interdependencies between teams, the utilization of the knowledge exchanged, and the external conditions surrounding KT will positively influence teams’ capability in terms of team and task familiarity. Although the mediums used for KT influence teams’ capability, the influence was found in negative direction. Areas of knowledge used and exchanged however have shown no influence on teams’ capability. The primary theoretical contribution of this research lies in the model that illustrates the interrelationships between analyst and software architect teams during non-collocated software architecture development, along with the influencing factors of KT over those teams’ capabilities. Taking team and task familiarity in viewing teams’ capability signifies a distinct approach that potentially inspires for greater contributions particularly in teams’ capability area of interest. In short, it provides better prospects for KT to occur between different functional teams that are non-collocated and rest a good understanding of how KT may influence their capability