Buying time in software development: how estimates become commitments?

TL;DR

This paper explores how software professionals use padding in estimates to manage commitments and human factors, revealing that estimation practices are deeply influenced by social pressures and the need to balance short- and long-term goals.

Contribution

It uncovers the human and social dynamics behind estimation padding and commitment formation in software development, emphasizing practical practices over purely technical estimation methods.

Findings

Padding serves as contingency buffer, task completion, or quality improvement.

Estimation padding helps buy time to balance commitments and technical debt.

Human and social factors influence estimation and commitment processes.

Abstract

Despite years of research for improving accuracy, software practitioners still face software estimation difficulties. Expert judgment has been the prevalent method used in industry, and researchers' focus on raising realism in estimates when using it seems not to be enough for the much-expected improvements. Instead of focusing on the estimation process's technicalities, we investigated the interaction of the establishment of commitments with customers and software estimation. By observing estimation sessions and interviewing software professionals from companies in varying contexts, we found that defensible estimates and padding of software estimates are crucial in converting estimates into commitments. Our findings show that software professionals use padding for three different reasons: contingency buffer, completing other tasks, or improving the overall quality of the product. The reasons to pad have a common theme: buying time to balance short- and long-term software development commitments, including the repayment of technical debt. Such a theme emerged from the human aspects of the interaction of estimation and the establishment of commitments: pressures and customers' conflicting short and long-term needs play silent and unrevealed roles in-between the technical activities. Therefore, our study contributes to untangling the underlying phenomena, showing how the practices used by software practitioners help to deal with the human and social context in which estimation is embedded.