Why change only one thing at a time?
Because afterwards you have to be able to read the result. Ship several edits at once and whatever happens next belongs to all of them and to none of them — the money is spent and the learning is gone.
One change keeps the question answerable: this edit, that asset, that window. Even when the honest answer is that nothing can be attributed yet, you know what the answer is about.
What one-at-a-time buys
- Attribution stays possible in principle: one change, one window, one reading.
- Reversal stays cheap: undoing one edit is an action, undoing five is an argument.
- The blast radius stays small, which is what makes approving it something you can do responsibly.
What it costs
Patience. A pilot that changes one thing cannot also fix everything else it noticed — those go into the next decision, where you choose to continue, hold or stop.
That trade is the pilot’s whole shape: what four weeks covers, and what it does not.
What this answer does not claim
- One-at-a-time makes attribution possible, not automatic. A single clean change can still end as cannot-be-attributed-yet, and that is reported in those words.
- This page argues a method. It does not claim the method moved any particular answer.