Instead of looking at overall save % in playoffs, or regular season, I think an individual game breakdown is needed. Basically, how many playoff games are won with sub .920 goaltending? I doubt many. How many of the teams that do manage to scrape out one or two series wins on the back of multiple sub .920 goaltending performance wins manage to win SCF? Remember it's OK to have an inconsistent goalie maybe whose save % is brought down by big losses, but then will bounce back with 2-4 strong performances, that will win you a playoff series. What I would not expect is a goalie who consistently puts up sub .920 performances, but his team keeps winning with strong goal support.
Looking at the Niemi example, who was one of the weakest of the 8 goalies TOML posted, and on arguably the best SCF winning team of the lot. 7 of his 16 wins were done with poor play, not just sub .920, but sub .900. Three of those were against the Flyers in the final, who had even worse goaltending though, which was kind of helpful for the Hawks. But the rest of his 9 wins he posted excellent #s. He had 6 losses, five of which he was under .900. One was a solid .917.
MAF in 2009 had 6 wins where he posted under .920. Not as bad as Niemi, only 2 were under .900, the other 4 were between .900 and .920.
So IMO elite goaltending does matter, a lot. Even goalies who posted less than elite stats were actually pretty elite in the majority of their wins, but weren't consistent and lost hard when they they did lose, dropping their stats. And also their teams were so good they managed to overcome a half dozen bad to mediocre starts and put up the extra goals needed to overcome.