In the "fireside chat" thread, we find WotCapparently asserted that GMs account for about 20% of the market, but a vast majority of the sales. From this we can glean a couple of points:
1) Beyond time as a resource for GMs, there's a monetary component as well.
2) If you fill a bag with 80 white marbles, and 20 red marbles, there are enough red marbles to go around in theory, but in practice if you grab any five marbles, there's a goodly chance they'll all be white.
#2 there is a problem - because you'll frequently have local GM shortage, but not an overall shortage. But, raising the number of GMs may not actually help - because then you'll tend to have areas where GMs who go underutilized, and drop out of GMing, and you end up back where you started.
Indeed, the probability of drawing 5 white marbles is very close to 31.9%, meaning about a third of all 5-person groups don't have a DM among them. (Assuming you're drawing without replacement. If you draw with replacement, it's just .8^5 = 0.32768 = 32.768%.) For 6-person (that is, hoping for "5+DM"), it's very close to 25.2% without replacement, or .8^6 = 0.262144 = 26.2144% with.
Given the sheer number of groups of gamers hoping a DM will be willing to pick up their premise on various forums I've been part of, I don't think there's too much risk of DMs "going to waste." (As I've said elsewhere, I spent a full year on multiple sites hoping to find a game across several systems, only to strike out.) Even if there were, however, there's a better solution.
Make more
pink marbles.
That is, make more marbles that can work as the red marble in a group of otherwise all white marbles, or a white marble if the group already has a red one. E.g., let's say we take your example, but we make it 70% white, 20% red, 10% pink. That cuts the chance of a 5-person group not having any DMs from the aforementioned ~31.9% to ~16.1%, nearly half (with replacement: 32.8% vs 16.8%.) For the 6-person group, it goes from ~25.2% to almost exactly 11% (with replacement: ~26.2% to ~11.8%), a reduction of
more than half. Despite only altering one-eighth of all "never DMs" into "sometimes DMs," you drop the proportion of potential-groups-lacking-DMs from one-in-three to one-in-six for five-player groups, and from one-in-four to one-in-
nine for six-player groups. That's a pretty major impact despite being only a small shift in the overall population.
The questions then become: What constitutes a pink marble? How do we encourage some white marbles to become pink ones? Will some red marbles also become pink marbles? How would either of these affect sales?
I'm of the opinion that well-designed, well-explained, well-balanced rules, which are made for a clear and useful purpose, are one of the best ways to encourage more pink marbles. Pink marbles are folks who have seen that it isn't
that hard to be a DM, if you can make a bit of time and become comfortable with improvisation and not
needing to plan everything to the nth degree. That it is possible to create accessible, well-designed tools (both in the game mechanics sense and the digital applications sense), and either sell them flatly at a reasonable price or make basic versions free and feature-rich versions reasonably priced.
Serious design, which pursues what
works under conditions of actual play
testing, and not just faux-playtest-as-marketing-gimmick, can get you a good portion of the way there.