When is a client change actually worth repricing?

I've been looking at how small service businesses handle existing clients on fixed fees.
One thing keeps coming up:
A client changes.
More transactions.
More emails.
More meetings.
A new service.
More revisions.
But not every change should trigger a price increase.
A temporary spike is different from a structural change.
A few extra hours one month is different from a permanently larger scope.
And "the client is taking more time" is often too vague to make a pricing decision confidently.
So I'm increasingly interested in the decision point before repricing:
What evidence is enough for you to say: "This has changed materially — it's time to review the price"?
Is it:
• a % increase in workload?
• additional services?
• transaction volume?
• hours consistently exceeding expectations?
• repeated scope creep?
• something else?
I'm particularly interested in how founders and service businesses actually make this decision in practice — not what the ideal process looks like on paper.