"Change management is slowing us down" is something I hear from almost every engineering team I work with, and they're usually right — not because change management itself is wrong, but because most implementations route every change through the same heavyweight approval process regardless of actual risk.
The Core Problem: Flat Risk Treatment
A configuration tweak with a trivial rollback and a database schema migration with no rollback path are different categories of risk. When both go through the same CAB approval cycle, the low-risk change waits unnecessarily, and reviewers spend their limited attention spread thin instead of focused where it actually matters.
The Fix: Genuine Risk-Based Tiers
Standard Changes
Pre-approved, well-understood, low-risk changes that don't need case-by-case approval at all — just logging. If a change has been executed successfully dozens of times with no incident, it shouldn't need a fresh approval every time.
Normal Changes
Genuinely novel or moderate-risk changes that warrant real review — but reviewed by people with the context to actually evaluate the specific risk, not a generic committee rubber-stamping a queue.
Emergency Changes
A clearly defined, pre-agreed fast path for incident remediation, so on-call engineers aren't stuck choosing between following process and resolving a P1 quickly.
The goal of change management isn't to slow everything down equally — it's to put scrutiny where risk actually lives and get out of the way everywhere else.
What This Looked Like in Practice
On programs I've run involving full project lifecycle management for transitions valued up to $5M, properly tiered change management was consistently what let teams move fast on routine changes while still protecting the genuinely risky ones — which is also a major contributor to delivering 20+ transitions with zero Sev-1 incidents post go-live.
If your engineering team thinks change management exists to slow them down, that's a signal your tiers aren't calibrated to actual risk — not a signal that change management itself is the wrong idea.