I second Prateek’s point. Great explanation. My colleague at the University of Pittsburgh studied a LEED Gold (less-energy-efficient) and a Net-Zero (more-energy-efficient + on-site PV) buildings in Pittsburgh, PA, using both approaches, i.e. the typical average approach vs. using hourly energy consumption and power generation data. He found that the typical LCA approach underestimates the impact of the less-efficient building and overestimated the impact of the net-zero building. It should be noted that these were both primarily office buildings, and as Prateek mentioned, the use type will also greatly affect the outcomes. It depends on the building and the grid. The full paper is here.