No one said you didn't but you also didn't express what your concerns about the web in this particular instance was.
Fair enough. And thanks by the way, I was having a tough time leaving this alone.
First I'll just point out the parts that "raised my hackles" as it were. You said that the flanges were "most important" which isn't true, every part of a beam design is important. The you said "most of the strength is in the flanges" which is also not true, most of the strength in a (I'm adding an assumption at this point but I think it is fair) economical beam design is in d. ntsqd said "Only in as far as it's needed to keep the upper and lower flanges equidistant;" well yeah, but a very important job it is!
Now to the issue: We are taking a beam that we should assume is designed to be economical that is configured as a cantilever from a given fulcrum. We will just assume that the load on the entire beam is uniform. Moving the fulcrum such that the cantilever section is lengthened will increase the moment on the cantilever and the shear with the highest shear values at the fulcrum. The web does not participate in carrying the moment much at all but does carry the shear stress. The web was sized for some shear stress and having moved the fulcrum we can exceed this. The web will crush/buckle, d decreases, the capacity of the beam decreases and stuff happens.
I didn't say there would be an issue because I have no idea. I believe that usually when an engineered beam fails the mode is torsional buckling of the compression flange. Part of the reason this is the usual failure mode is because most beam designs have stout webs due it being really easy and not very costly to over engineer and using web stiffness to counter torsional stress (which creates a web with way more material needed for the shear stresses). I wouldn't expect this in our case because I believe the compression flange is fully restrained through the skin stuff that is the bottom of the overhang and being fully boxed at the ends so it is probably over engineered for torsional stress. (heh, the webs of the cantilevers are flanges to a big horizontal beam by the way.) I would expect our venture to cause web crushing with subsequent yield failure in the compression flange.
BUT, I don't know. I know nothing about the beam design they've used.
[edit: I should have kept reading...]
Adding d is always the choice when allowed given yield strength of materials and resistance to torsional buckling. Web is cheap.
The compression flange is (I believe anyway) fully restrained so torsion probably isn't going to be the first thing we run into.
I was thinking it was a box so I really thought Pods8's instinct was safe enough and leaving the web discussion alone was fine. Now that I see that it is a C things are different. If there is wood inside the C then this can act like a beam inside a beam. In fact, if the C doesn't have this then we can add it compensate for the delta shear/moment being aware that we move load concentrations so we need to go far enough that shear/moment stresses are at least as low as they are with the original. (which is further than the original fulcrum by a bit.)
[edit #2]
So, I left this and went a watch a bit of TV and suddenly it occurred to me that it is really common to add web thickness and doublers in wing spar designs - like the wood beam I mentioned above. Rivet some angle along both flanges to the flange and then rivet a doubler for the web across the angle(so, original web, angle, web doubler). Oh, all of this on the inside of the C.