One concrete data point for the thread. If you are going to change something, change one thing and give it long enough to express itself. Four weeks is the usual minimum for anything pharmacological on this board, and two weeks of data has told you almost nothing.

One thing that is still open after LibrarianMeg’s answer:

Whether the fasting requirement is as strict in practice as the label implies, and what people actually see when they get it wrong?

DataDave said:

If you are going to change something, change one thing and give it long enough to express itself.

Coming at DataDave’s question from a different direction. Start from the measurement rather than the conclusion. Almost every disagreement here turns out to be two people measuring different things and comparing the numbers anyway.

Closing the loop on my own question.

Follow-up: moving the tablet to the moment I wake, before anything else, was the whole fix. The dose never changed.

jim_asheville said:

Start from the measurement rather than the conclusion.

jim_asheville has the substance of this right. The condition it depends on is worth stating. The absorption variability is real, but it partly averages out over weeks — the steady-state trough is less erratic than any single day would suggest. Where it bites is in the first fortnight, when people conclude the tablet does nothing.