The point of a log is not the record; it is that writing "no" in the last column three times in a row is a conversation with the client rather than a discovery at the end of the project.
A log with a row per change: what was asked, when, its effect on dates and hours, what it was priced at, and what was decided.
The absorbed rows are the point. Changes you agreed to for free are invisible everywhere else and are usually where the margin went.
Log the request before assessing it, including the ones you will obviously absorb. The record of the request is what makes the pattern visible later.
Price everything, even the absorbed ones. An absorbed change with no value attached teaches nothing.
It will not stop scope creep. It makes it countable, which is the precondition for doing anything about it.
It also cannot price a change you have not scoped.
At the end of a project, total the absorbed column. That figure is the conversation to have — with the client if it is large, or internally about how requests are handled.
A client whose absorbed changes exceed the project's margin is not a profitable client, however good the relationship feels.
Yes — the point is the total, and the total is made of tiny things.
The project owner, updated as requests arrive.
Yes, no email required.
Half an hour on your own figures, and an honest answer about the parts Treepie does not improve.