From this (https://www.sciencedirect.com/science/article/pii/S095965261...), they estimate a scaled up process requiring ~50kg of hydrogen to generate 1000kg of steel (which would be about 1500kg of iron ore).

50kg of hydrogen should be equivalent to like ~600kg of ammonia (if I did my math right). So there's a substantial mass difference in one way shipping of 600kg of ammonia versus two way shipping of ~1000kg of solids to consider. In addition, since the tanker is empty on the return trip, if we were willing to play the travel time game, we could reduce tanker return speeds to try to compensate for running faster while carrying.

So I don't think it's completely cut and dry that shipping the iron ore/steel makes more sense.