What You Learn When You Actually Track Postmortem Follow-Through

Everyone writes action items.
Almost no one follows through.

You know the routine. Incident happens. A team scrambles to recover. A postmortem is created. A few action items are scribbled in at the end to make it look like progress was made. Then everyone moves on. Until the next outage, which looks suspiciously like the last one.

What happens to those action items? Most of the time, absolutely nothing.

When you actually track follow-through, the picture is eye-opening. And not in a good way.

Action Items Are Easy to Write. Hard to Own.

Postmortem action items often suffer from a few fatal flaws:

• They are too vague. "Improve alerting" or "clean up tech debt" means nothing without context.
• They have no owner. No one knows who is supposed to do it.
• They have no deadline. Which means they never get prioritized.
• They do not connect to risk. So teams treat them as chores instead of resilience investments.

You would be surprised how many companies with dozens of incidents per year can only point to a handful of closed actions. That is not a documentation problem. It is a feedback loop problem.

Tracking Follow-Through Reveals Organizational Truths

Once you start tracking action items properly, across teams, over time, a few patterns jump out.

Teams under pressure skip cleanup

If action items are not part of sprint planning, they fall off the radar. Especially for high-velocity teams.

Repeated failures are often ignored

Same service. Same class of issue. Same fix proposed multiple times. No closure.

Orgs reward speed, not depth

Fast resolution is celebrated. Deep remediation is optional. Until the incident repeats and escalates.

Engineering memory is short-lived

As team members rotate or leave, past lessons disappear. The action item might be closed in Jira, but the knowledge is lost in practice.

A Follow-Through Rate Below 50 Percent Means Your RCA Is Broken

The purpose of postmortems is not to tick a box. It is to drive change. If your follow-through rate is below 50 percent, you are not learning. You are writing reports.

Teams that actually close out their postmortem actions do a few things differently:

• They treat action items as first-class engineering work.
• They make action items visible and traceable.
• They use data to understand which types of incidents are generating the most ignored work.

This is not hard to do. But it requires intent. And a system to back it up.

You Can't Improve What You Don't Track

If you are not measuring follow-through, you are guessing at progress. You might have great writeups. You might have smart engineers. But you do not have a learning culture.

Start simple:

• Log every action item
• Track ownership and closure rate
• Review patterns monthly or quarterly
• Connect actions to outcomes

Better yet, let a system track it for you; one that knows when the incident happened, what was promised, and whether the change ever landed.

This Is Why We Built COEhub

At COEhub, we believe operational learning should not rely on memory or motivation. It should be automatic, structured, and visible.

COEhub tracks postmortem actions across incidents, surfaces those that are ignored, and helps leaders prioritize the ones that matter. It is not just about writing better postmortems. It is about making sure they lead to something real.

If you are serious about resilience, you have to close the loop.

Start tracking follow-through. And find out what your org is really learning.

Learn more at COEhub