Jim Green's comment might have been a result of InSight launching at the very beginning of the launch window. If you launch from a given point on Earth's surface into a polar or near-polar orbit, the orientation of the orbit's plane depends on where the launch site was, considering Earth's rotation, when launch occurred. Every hour that launch site rotates 15° with respect to the Earth-Sun direction, so the resulting orbit plane rotates by 15° as well.
For now assume a perfectly polar orbit for the InSight parking orbit. It wasn't quite polar, but it was close enough that this approximation will aid understanding.
For maximum efficiency the V∞ vector of the intended Mars transfer trajectory should be close to parallel to Earth's orbital velocity vector in right ascension, and somewhere close to the ecliptic plane in declination. The ideal launch and parking orbit will have the intended V∞ vector in the orbit plane so the injection maneuver needs no out-of-plane component.
But that means you have a launch window of only instantaneous duration, not good for keeping a schedule.
InSight had a 2-hour launch window. If it were centered on that ideal launch moment, and they launch at the opening of the 2-hr window, then they are launching into an orbit plane that is 15° "early": Vandenburg is 15° farther west than it should be for the ideal orbit orientation, and the resulting orbit plane is offset 15° clockwise (as seen from the north). As it comes over the north pole for the trans-Mars injection (TMI) maneuver, it needs a significant out-of-plane component to that maneuver to meet the required departure V∞. And indeed, it would be to the left, if you're sitting upright on the Centaur, Jim's "mean left"!
Had InSight launched at the end of the launch window, 2 hours of planetary rotation would have moved Vandenburg 15 east of the ideal spot, and the out-of-plane component of the TMI burn would have been to the right — a "mean right".