Yes, I don't doubt at the
federal level that parallel construction works better than it does at the local level, because the feds have jurisdiction over bigger crimes. Drug crimes, involving multinational criminal organizations, generally involve many participants and transactions, and you only need to break 1 participant in a conspiracy to unravel the whole thing. (And more importantly, due to the many overlapping crimes being committed toward the common end, there are simply more
options for pursuing parallel construction or even just pursuing different crimes.)
But at the local level where most crimes occur, parallel construction is a fantasy for the overwhelming majority of cases. Cops would absolutely love if they could just parallel their way around tainted evidence, but in most cases, there's only one crime, and only one trail of evidence, and there's no way to parallel around that.
(And I speak from experience; I was a public defender before I went into tax. I've seen many more failed attempts at parallel construction than successful attempts.)