A ranking is never truly final, because the index is continuously re-evaluated, so the better question is when a ranking has settled into a stable range. The premise that a page reaches a permanent fixed position is the part to drop first. Rankings are an ongoing judgment that can shift whenever the page, its competitors, or the algorithm changes. What you can identify is not a final number but a settled state: a page that has held a consistent range across multiple crawl and update cycles with no clear upward or downward trend.
The reframe matters because “final” sets the wrong expectation and leads to the wrong reactions. If you believe a position is fixed, every small wobble looks like a problem and every dip looks like a loss of something permanent. Once you treat ranking as a continuously re-assessed estimate, normal day-to-day movement stops reading as failure. The signal you actually want is stability over time, not a date on which the page supposedly locked in, because no such date exists.
So the practical pivot is to judge settled by behavior across cycles. A page is settled when it has bounced within a narrow band, say a stable handful of positions, through several crawls and at least one or two broader update windows, without trending in either direction. A page is still settling when it is climbing, sliding, or swinging widely, which is common in the first weeks after publishing while Google is still testing its placement. Stability across cycles is the evidence; a single good week or a calendar milestone is not. Treat the specific cycle counts as a working guide rather than a fixed rule, since they vary by site and query.
To act on this, stop waiting for a position to become permanent and start tracking whether it holds a stable range over multiple crawl and update cycles. When it does, treat it as settled and stop reacting to minor noise. When it is still moving, keep observing rather than intervening, and remember that even a settled ranking can shift later, so monitoring never fully ends.