He has direct influence over some of the Okudagrams on TNG-R (which override the previous canon), while Orci obviously used the conjectures to come up with dates like 2230 or 2233 in STXI. Before that, TOS-R had confirmed Greg Jein's conjectural registry number scheme, also adopted by Mike Okuda.
I'm not saying his and Rick Sternbach's works are canon, only that it makes no sense to ignore them, since you never know how much of that will eventually become canon. Those works are that close to canon-makers, aside from their obvious influence on licensed publications.
Why adhere to one interpretation when you're writing a tie-in novel and are bound by Okuda's conjectures, another when you're not? Star Trek should be more integrated than that. So we start with the official interpretation and add the Enterprise
-class bit, which is also supported by the simulator door label. There is no need to ignore evidence, canon, near-canon or only licensed.