Software projects often operate in high-uncertainty scenarios. Best approach may often only be "obvious" in hindsight. It's not irrational to do it in parallel/ in several teams - in fact, I'm surprised it's not happening more often.