When Continuous Delivery Is Not an Option: Practical Paths to Continuous Engineering in Complex Organizations

TL;DR

The paper investigates why continuous software engineering (CSE) is not universally feasible in complex, regulated organizations by analyzing four industrial cases across automation, automotive, retail, and chemicals. It applies and extends the CSE Industry Readiness Model to assess current and potential adoption levels through expert interviews and narrative synthesis, revealing that organizational readiness and cross-organizational dependencies strongly shape adoption. The findings argue that meaningful internal improvements are achievable even when full end-to-end CSE is blocked by market, regulatory, or ecosystem constraints, and they propose an updated readiness framework with distinct internal/external feedback loops and improved guidance for setting realistic goals. This work provides empirically grounded, practically applicable guidance for navigating constrained CSE transformations in real-world, regulated environments.

Abstract

Purpose: Continuous Software Engineering (CSE) promises improved efficiency, quality, and responsiveness in software-intensive organizations. However, fully adopting CSE is often constrained by complex products, legacy systems, organizational inertia, and regulatory requirements. In this paper, we examine four industrial cases from the automation, automotive, retail, and chemical sectors to explore how such constraints shape CSE adoption in practice. Methods: We apply and extend a previously proposed CSE Industry Readiness Model to assess the current and potential levels of adoption in each case. Through expert interviews and narrative synthesis, we identify common driving forces and adoption barriers, including organizational preparedness, cross-organizational dependencies, and limited customer demand for continuous delivery. Results: Based on our findings, we propose an updated readiness model that introduces additional levels of internal and external feedback, distinguishes market- and organization-facing constraints, and better guides practitioners in setting realistic CSE adoption goals. Conclusions: Our results highlight that while full end-to-end CSE adoption may not always be feasible, meaningful internal improvements are still possible and beneficial. This study provides empirically grounded guidance for organizations navigating partial or constrained CSE transformations.

Figures (3)

• Figure 1: The conceptual mode of continuous software engineering (adopted from Klotins et al. klotins2022towards)
• Figure 2: Industry readiness model (adopted from Klotins et al. klotins2023continuous)
• Figure 3: The updated industry readiness model.