The Thriving In Engineering Digest: 13-25

A curated week.

The Thriving In Engineering Digest is live!

Every week, I filter through the trending topics and handpick a selection to bring you up-to-date on the latest news in AI, team dynamics, managing up, launching products, and more. It’s your one-stop shop for all the sharpest ideas in tech leadership.

Full access to this leadership shortcut is for paying subscribers only – and if you’re one of them, I’m glad you’re here.

Let’s get into it.

1 - Why 'Never Start with 100%': A Guide to Progressive Software Delivery

Here, Jason Yip delves into progressive delivery techniques for software products, emphasizing methods to reduce downtime and the impact of mistakes on users. The post critiques the “Big Bang Release” approach, where products are released to 100% of users, making mistakes highly impactful.

Yip explores the “Blue-Green” method, which allows for quick rollbacks by maintaining parallel environments, though it's not without challenges, too. The “Parallel Release” validates changes using live data without customer exposure, a strategy often used for machine learning models.

“Progressive Rollout,” also known as the Canary Release, and “Targeted Rollout” involve incrementally releasing software to specific customer subsets to limit risk. The “Guarded Release” option automatically adjusts rollout based on performance metrics, minimizing negative impacts with minimal human intervention.

Main Takeaway:

Start software deployments with smaller user segments to minimize risk, validate changes, and enhance learning.

Why It Matters:

For engineering leaders, adopting progressive delivery methods reduces customer disruption, improves feedback loops, and enhances deployment confidence, ultimately contributing to more resilient software delivery practices.

Read This If You’re Interested In:

• Exploring alternatives to Big Bang Releases

• Understanding Blue-Green deployments

• Experimenting with Progressive and Targeted Rollouts

• Learning from real-world usage data during releases

• Balancing innovation with risk management