Accurate task estimation is critical in any project management. Effective estimation allows teams to allocate resources efficiently, set realistic expectations, and manage project risks. Therefore, understanding the factors influencing the estimate’s accuracy is essential to improve project management practices and achieve better results. The study aimed to identify the optimal task time ranges to minimize estimation errors. The analysis focused on task durations estimated based on Story Points, examining whether shorter times lead to more accurate estimates. In addition, the study considers the potential improvement in the estimate’s accuracy with the project’s progress. Based on a publicly available dataset, the analysis examined estimation errors in different projects and issues that lasted less than an hour to more than eight days. The results show that tasks estimated to take 4 h to 1 day and 1–2 days had the smallest estimation errors. The latter was confirmed for user stories, which do not need to be split into shorter tasks. Tasks estimated at 0.5 h were consistently underestimated, and those with durations greater than eight days were overestimated, leading to more significant errors. Furthermore, the study found that the estimate accuracy did not improve as the project advanced, possibly due to unpredictable tasks such as bug fixing.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Optimizing Task Estimation Accuracy in Agile Projects: A Data-Driven Study

  • Bogumiła Hnatkowska,
  • Radosław Grębski

摘要

Accurate task estimation is critical in any project management. Effective estimation allows teams to allocate resources efficiently, set realistic expectations, and manage project risks. Therefore, understanding the factors influencing the estimate’s accuracy is essential to improve project management practices and achieve better results. The study aimed to identify the optimal task time ranges to minimize estimation errors. The analysis focused on task durations estimated based on Story Points, examining whether shorter times lead to more accurate estimates. In addition, the study considers the potential improvement in the estimate’s accuracy with the project’s progress. Based on a publicly available dataset, the analysis examined estimation errors in different projects and issues that lasted less than an hour to more than eight days. The results show that tasks estimated to take 4 h to 1 day and 1–2 days had the smallest estimation errors. The latter was confirmed for user stories, which do not need to be split into shorter tasks. Tasks estimated at 0.5 h were consistently underestimated, and those with durations greater than eight days were overestimated, leading to more significant errors. Furthermore, the study found that the estimate accuracy did not improve as the project advanced, possibly due to unpredictable tasks such as bug fixing.