> Isn't trying a plausible sociological intervention then measuring its effectiveness a win regardless of how effective it ended up being

No, paying $3B for information of dubious value and which was pretty obvious ahead of time is not a win.