← How we think

An org changelog that humans can read

What did the team ship this week? Not commits. Not tickets. Features.

Every engineering leader has the same problem: they want to know what their teams are delivering, but the tools they have show the wrong thing. Git logs show commits. Jira shows ticket state changes. CI dashboards show pipeline runs. None of these answer the question they are actually asking: what changed for our users this week?

The gap between engineering artifacts and business outcomes is where misunderstanding lives. A manager sees 47 commits and has no idea if that was one feature or a week of refactoring. A team lead sees 12 closed tickets and cannot tell if they were critical fixes or backlog grooming. The information exists, but it is trapped in a format designed for engineers, not for the people who need to communicate progress to the rest of the organization.

The org changelog

gittan maintains an org-level changelog — a feed of what each team has shipped, written in plain language. Not commit messages. Not ticket IDs. Descriptions of features, fixes, and changes that a non-engineer can understand.

When a team pushes a change, the pipeline passes, and code lands on main, the team can attach a changelog entry. A short description of what changed and why it matters. "Users can now export invoices as PDF." "Search results load 3x faster on mobile." "Fixed a bug where notifications were sent twice after password reset."

These entries accumulate into a timeline per team. At the org level, every team's timeline merges into a single feed. A CTO opening gittan on Monday morning sees what every team shipped last week — in language they can forward to the board, paste into a stakeholder update, or use in a quarterly review.

Why this is not a commit log

Commit messages are for engineers. They describe what changed in the code: "refactor auth middleware to use token rotation" or "bump openssl to 3.2.1". These are useful in a git log. They are meaningless in a leadership update.

The changelog is a separate artifact. It is written by the team, for the organization. Not every push needs an entry — a dependency bump or a typo fix is not worth communicating. But when the team ships something that matters to users, to stakeholders, or to other teams, they write a line about it. The discipline of writing that line also forces the team to think about what they are delivering in terms of outcomes, not outputs.

Visibility without surveillance

The common alternative is status meetings. Standups, weekly syncs, sprint reviews, all-hands demos — ceremonies where teams report what they have done to an audience that could have read it instead. These meetings exist because the information is not available anywhere else in a consumable format.

The org changelog makes those meetings optional. The information is there, updated continuously, in a format anyone can read. A product manager can see what the backend team shipped without joining their standup. A VP of Engineering can see what all teams shipped without scheduling a round of 1:1s. The information flows without the ceremony.

This is not surveillance. The changelog does not track who committed what, how many lines were changed, or how many hours someone worked. It tracks what the team — as a unit — delivered. The team writes the entries. The team decides what is worth communicating. The team controls the narrative.

A shared language

One of the hardest problems in engineering organizations is translating between engineering language and business language. Engineers think in systems, APIs, and deployments. Business thinks in features, customers, and revenue. The changelog is a bridge.

When an engineer writes "users can now filter dashboards by date range," both the team and the stakeholder understand what shipped. There is no translation layer, no project manager writing a separate status report, no game of telephone between the person who wrote the code and the person who reports to the board.

Over time, the changelog becomes the canonical record of what the organization has built. Not in terms of code, but in terms of capability. That is the thing that actually matters.