Research — Pillar 02
People, Teams & Organisations
Software engineering is a human activity. Even now. The most consequential variables in delivery outcomes are not technical — they are organisational, interpersonal, and psychological.
Three bodies of empirical work fall under this pillar. They address different questions, but share a common finding: the conditions surrounding the work matter as much as the work itself. Agile frameworks, remote work arrangements, and team composition all interact with human variables in ways that technical analysis alone cannot capture.
Agile Transformation
Agile is everywhere. Whether it works well is a different question, and one that depends heavily on context.
A mixed-methods study of large-scale agile transformations identified four factors that consistently predicted success: top management commitment, Product Owner involvement, Scrum Master leadership, and developer skills, both technical and social. Without all four, transformations stalled. The presence of a recognised agile framework was not among the predictors.
Teams that tailored frameworks to their own context outperformed those that adopted any framework as prescribed. Greater team autonomy correlated with both higher satisfaction and higher productivity. This is not an argument against frameworks — it is an argument against treating them as fixed specifications.
Psychological safety directly improved team effectiveness in agile settings. Teams where members felt safe to raise concerns and challenge decisions showed measurably better outcomes across team sizes and organisational contexts. Proactive conflict management prevented disruptions; unmanaged conflict eroded the benefits of autonomy.
The practical implication: a successful agile transformation is an organisational change programme, not a process installation. The human and structural conditions have to be built alongside the ceremonies and tooling.
Future of Work and Digital Resilience
The COVID-19 pandemic created an unplanned natural experiment in remote software engineering. The findings challenge several popular assumptions from both the optimistic and pessimistic camps.
Software engineers adapted to remote work faster than expected. They were already comfortable with digital tools and continued to deliver at comparable levels. The transition itself was not the primary source of disruption.
The strongest predictor of well-being across both measurement periods was the quality of social contacts, not their frequency. Engineers with high-quality social interactions reported better well-being regardless of how often they interacted. This has implications for how hybrid and remote teams should be structured: fewer, higher-quality touchpoints outperform high-frequency low-quality contact.
Boredom and distraction reduced productivity measurably. Engineers who reported high levels of either showed lower output, pointing to a structural issue: without a well-organised work environment, remote productivity suffered regardless of technical competence.
People adapted over time. As lockdown conditions persisted, the relationship between predictor variables and productivity weakened. Engineers found their own coping strategies. The initial disruption faded even when the conditions did not change. Gender differences were minimal throughout.
Diversity and Inclusion
Diversity in software teams has measurable effects on team dynamics. Those effects are not uniformly positive, and research that treats them as such is not useful to teams trying to manage them.
A study of 1,118 participants across 161 agile software teams produced four specific findings. Age diversity improved team effectiveness. Gender diversity increased relational conflict — and introduced a wider range of perspectives. Both effects coexisted, which is precisely why conflict management strategies are not optional when building gender-diverse teams.
Psychological safety mediated both effects. When team members felt safe to speak up, diverse teams performed better. When they did not, diversity created friction without corresponding performance gains. Psychological safety is not a secondary condition — it is the mechanism through which diversity converts into performance.
Role and cultural diversity showed no significant effect on team effectiveness or conflict in this sample. A separate study using the HEXACO personality model examined gender differences in personality traits among software engineers specifically. Meaningful differences were found across several dimensions, with practical implications for team design.