Broad match plus tROAS - has anyone switched from phrase and NOT regretted it?

Every rep call for the last eighteen months has ended with the same suggestion. Go broad, let the bidding sort it out, stop fighting the machine. And every time I have tried it on a client account the search terms report fills up with things I would never have paid for... I am willing to accept that I am the problem here. The theory is sound enough: if the conversion data is good and the target is honest, broad plus tROAS should find pockets that phrase never reaches. I have just never once seen it beat a well-built phrase account over a full quarter, and I have tried it maybe six times. So the question is for anyone who made the switch and stayed switched. Not the first two weeks, not the case study your rep showed you. A quarter later, was it actually better, and what did you have to have in place first? My suspicion is that it needs far more conversion volume than most accounts have and nobody wants to say so out loud.