I like this move because stats can be used not only to evaluate players, but also to help them improve in specific areas.
I think keeping Nonis makes sense because even though he made the Clarkson deal he is a decent GM. Arguably the Clarkson deal was made at the behest of Carlyle and his coaching style.
I think Shanahan keeping Nonis makes more sense then keeping Carlyle. That said, if you tell a coach that he has to coach a different way and use tools they are unfamiliar with, and they are comfortable doing that, then maybe you keep him. The collapse at the end of last season could have been a wake-up call for Carlyle to change his game, with an extra push in that direction by Shanahan.
Dubas should be an interesting addition to the Leaf's front office, but his performance will be really hard to judge unless we get media leaks saying he did this or he did that.
We should all be looking for the Leafs to continually build and improve for the short and long term. Instead we want change for change's sake.
Here is the amount of time each of these people have spent in their current roles.
Shanahan - 3 months, 15 days
Nonis - 1 year, 6 months, 17 days
Dubas - 4 days
Carlyle - 2 years, 4 months, 24 days
It's pretty hard to effect immediate changes in an organization. Dubas did it as GM with the Sault Ste. Marie Greyhounds in 3 seasons.
After 5 years, then we should have a better idea of how good or bad these people are. But we have way too small of a sample size with them all working together to be able to say anything about how this will work out.