Effort Estimation in Agile Environments: An Empirical Approach to Understanding Its Application in the Software Industry

Abstract

Effort estimation remains a central challenge in agile software development, where teams must balance accuracy and adaptability within dynamic delivery cycles. While various estimation techniques are proposed in the literature, there is limited empirical evidence on how these practices are applied in realworld settings. This study presents an empirical investigation based on a survey designed to explore current estimation practices in agile environments. The survey instrument was informed by a prior Systematic Literature Review (SLR), which identified key strategies, techniques, and research gaps in agile effort estimation. The field study targeted professionals actively involved in planning and estimation tasks across agile software teams in Argentina, with additional responses from neighboring countries. Participants included developers, product owners, and scrum masters, allowing for a multi-perspective analysis of how estimation is practiced across different roles and levels of experience. The survey focused on tools used, estimation techniques adopted, and the perceived internal and external factors influencing estimation accuracy. Findings reveal a coexistence of quantitative and qualitative approaches, along with inconsistencies in the integration of estimation techniques into agile workflows. Key challenges identified include selecting appropriate metrics, dealing with incomplete requirements, managing team experience variability, and addressing stakeholder-driven pressures. The study highlights that the effectiveness of estimation practices is not solely dependent on tools, but on the effort invested in interpreting and contextualizing estimation data. This research contributes to the understanding of estimation in agile settings and offers insights for refining estimation strategies in uncertain and evolving development environments.