Getting Feedback from Test-Driven Development and Testing in Production

2026-02-05 · Source: InfoQ · Field: Technology & Digital — Software Development & Engineering · Depth: Advanced, short

Summary

Ola Hast and Asgaut Mjølne Söderbom presented their approach to continuous delivery at QCon London, emphasizing a strategy that relies heavily on unit and integration tests, eschewing traditional end-to-end tests. Their methodology integrates Test-Driven Development (TDD), pair programming, and robust design principles to facilitate frequent, small deployments. They advocate for "testing in production" to obtain high-quality feedback, mitigating risks through feature toggles. This decade-long practice has led to a system where test environments are largely bypassed, and fast feedback loops from both tests and production drive continuous improvement and design refinement. They highlight TDD primarily as a design tool, with testing benefits being a positive side effect.

Key takeaway

For software engineering teams aiming to accelerate continuous delivery and improve code quality, you should consider shifting your focus from extensive end-to-end testing and dedicated test environments. Instead, embrace a strategy centered on Test-Driven Development, pair programming, and "testing in production" with feature toggles. This approach provides faster, higher-quality feedback, enabling quicker iteration and more resilient systems, ultimately driving better design decisions.

Key insights

Prioritize fast feedback loops and robust design over extensive test environments and end-to-end tests.

Principles

• Trust unit and integration tests.
• Test in production for best feedback.
• TDD is a design tool.

Method

Implement TDD with pair/mob programming, deploy small changes frequently with feature toggles, and use production for primary feedback, minimizing reliance on staging environments.

In practice

• Use feature toggles for new features.
• Deploy small code changes often.
• Focus on unit and integration tests.

Topics

• Test-Driven Development
• Testing in Production
• Pair Programming
• Continuous Delivery
• Software Design

Best for: Software Engineer, MLOps Engineer, DevOps Engineer

Related on AIssential

• Approval Testing vs TDD — The core debate is whether approval tests, despite their utility, truly align with TDD's design-first principles.
• Building production-ready LLM-powered applications @ Scale By the Bay — Transitioning LLM prototypes to production requires a pragmatic, product-focused approach centered on continuous evaluation and user outcomes.
• Continuous Delivery in a World of Constant Change • Abby Bangser & Dave Farley • GOTO 2025 — Continuous Delivery's incremental, verifiable steps are foundational for robust software development, especially with AI.
• The hidden pearls of TDD — TDD is a foundational discipline that enhances software quality, design, and AI agent effectiveness through structured feedback and complexity prevention.
• Testing Designs: Useful or Waste of Time? — Hands-on testing of design alternatives reveals optimal solutions that pure theoretical analysis might miss.
• From TDD to EDD: How engineering paradigms evolved in the age of AI — GenAI demands Evaluation-Driven Development (EDD) to systematically validate behavioral changes across multiple dimensions before shipping.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by InfoQ.