Why Severity Labels Stop Working
Severity labels lose value when criteria are vague, unevenly applied or never updated. Tie each level to business impact and recalibrate regularly.
Better Software Team
Blog
Insights
Articles on product strategy, engineering principles, and building software that lasts.
How to Test Non-Deterministic AI Features
How to test AI features when outputs are non-deterministic, with measurable criteria, eval sets, pass rates and graders, plus the mistakes teams make most.
Better Software Team
Planning Capacity for Production Work
Incident response, support and bug fixes consume engineering capacity that plans often omit. Measure that load, add a production buffer and cut its root causes.
Better Software Team
When a Landing Page Is a Better MVP
When a landing page MVP beats working software, how to run one step by step, how the two compare, and the common mistakes that make demand tests misleading.
Better Software Team
What Moving Fast Really Costs Engineering
Moving fast without engineering discipline builds technical debt, raises incident rates and slows future work. See which standards and metrics restore speed.
Better Software Team
Finding Silent Background Job Failures
A job that reports success can still skip its work. Check status, logs, payloads, external calls and database state, then monitor outcomes, not completion.
Better Software Team
Why the Same Incident Keeps Returning
Recurring incidents persist when fixes address symptoms instead of root causes. A structured investigation finds the shared preconditions and closes them.
Better Software Team
When Server-Side Rendering Matters
When server-side rendering matters for a product, how it compares with client-side and static rendering, and a page-by-page way to choose without overspending.
Better Software Team
Building a Failure Model From Partial Data
Turn logs, metrics, traces and anecdotes into a testable hypothesis when production data is incomplete, then validate it, narrow the scope and prevent repeats.
Better Software Team
A Refactor or Rewrite Assessment Checklist
Refactor localized technical debt and rewrite only for fundamental architectural limits. This eight-step assessment compares cost, risk and business impact.
Better Software Team
Spotting Reliability Erosion Over Time
Detect slow reliability decline from SLI trends using baselines, tail latency, segmentation, deployment correlation, dependency checks and load tests.
Better Software Team
CI/CD That Catches Bugs Before Release
Build CI/CD pipelines that fail on defects, with quality gates, layered automated tests, static analysis, security scans, staged deployments and rollbacks.
Better Software Team
