Universal testing and full identification, isolation, and clearance of all patients across multiple checkpoints would result in the fewest number subject to confinement, confined in the least broad of areas for the shortest periods of time.
OTOH if insignificant testing has been achieved, lockdown itself will need to be inversely universal and long-lasting.
There is an exponential component involved, so the effect will be felt sooner than you think.
Theoretically either way, it's possible no one will be able to be unconfined across multiple checkpoints without a real-time valid negative test or credential.
That decision could be the same either way.
But the fewer and less reliable the tests, the fewer the number of citizens that can be expected to be allowed to move about with some freedom.
After pandemic stress exceeds the magnitude where there can be no treatment resources for additional patients anyway, movement will certainly not be relaxed, and those who might have had a chance with earlier & more thorough action & testing will be left to expire undertested in confinement.
Not every scenario is that uplifting.
Due to the excess rate of confinement resulting from inadequate testing, the greater stress on the marginal economic needs of a large contingent of the populace may result in a need for lockdowns to become oppressive enough to theoretically halt the spread of all possible crime in addition to disease.
Depends on how serious it gets, if it's bad enough you could get unrulyness or looting like there is during some disasters regardless.
People will continue to get the Flu, Colds, HIV / AIDs even throughout this whole shebang. We need to know COVID19 cases specifically, to know if our lockdown methodology is being effective.
Once the growth is no longer exponential, measures can be rolled back slowly, especially if they have access to enough tests, masks and respirators.
The earlier action is taken, the cheaper it will be to stop this. Every week of delay can increase casualty rates by a factor of 10. A month of delay, can mean the difference between 10000 and 10 million dead.
Also, economic costs, both in direct treatment cost as well as the disruption to the productive side of the economy will rise sharply the longer we wait. (Well, at least if we don't see the death of sick and elderly as an economic benefit.)