Rethinking Dev-Test-Staging-Production Pipelines for Safety

For years, we convinced ourselves that Dev-Test-Staging-Production pipelines kept us safe. If we just had one more layer of validation and one more environment to test in, we’d catch the problems before they hit production.

Except… that was a lie.

Staging environments aren’t production. They’re expensive illusions of safety. The data is different. The scale is different. The behaviours are different. And when failures inevitably show up in production anyway, teams scramble, wondering why their extensive pre-release testing didn’t catch the issue.

Modern software delivery isn’t about pretending production doesn’t exist; it’s about embracing it. Audience-based deployment (sometimes called ring-based deployment) is the smarter approach: releasing features to small, targeted user groups first, monitoring real-world behaviour, and expanding based on what the data tells us.

You don’t need more pre-production gates. You need faster feedback in production.

Are you still clinging to staging environments? What’s stopping you from moving forward?