Positivity rate has always been a flawed statistic since the people who are tested are self-selecting (or other-deselected when told not to get a test unless you have symptoms).
That's why many here have always looked to other stats, such as hospitalization and deaths.
Tho, the positivity rate hasn't been too far off. In general, we see the hospitalization and death rates conform to positivity rate two to four weeks later, as is expected.
In the end, the timing of 'when things get back to normal' is when there are very few deaths and hospitalizations. Then, contact tracing can theoretically contain new cases.