If we're talking 5e, and we're assuming the bulk of sales are going to be in the form of DDI subscriptions, then they don't need to bother.
The game should support a basic set of monsters from the get-go, including all the classics. The 4e Monster Manual is actually not a bad starting point; the 3e Monster Manual would probably be a better set. (But that's a quibble - it doesn't matter all that much.)
Then, provide a quick and easy toolkit for creating new monsters, powerful enough so that any DM who requires an unsupported monster for their game is able to convert it by hand easily. (They could call it something like... "Monster Builder".)
But here's the clever bit: When a DM stats up a creature in this manner, give them the option (and try to encourage them) to 'open' their work for others to use. Over time, anything that anyone actually uses will be converted without WotC having to pay to do it. (For extra credit, they should employ someone to go through these various conversions, and select an 'official' version for later use.)
And there it is: minimal investment required, support for whatever is needed from old editions, and a DDI that continually increases in value.
Repeat the same feat with an "Item Builder" and a "Power Builder" (perhaps even a "Class Builder" and/or an "Adventure Builder"?), provide rating/errata tools for DMs to use, and you get all the theoretical benefits of the "open gaming" movement while keeping things in-house. The DDI continually increases in value, and the investment required is pretty minimal (you'll need to employ people to identify the 'official' items and powers that are useable for organised play and the like, and probably need to prune/nerf the fan-made versions to counter munchkinism).
The only people who lose out are the Design & Development teams - such a project kills any value in WotC producing new monsters, items or spells. And so, if they do their job right and create toolkits that are powerful enough, they've basically worked themselves out of a job. Oops.
(Oh, and once again: the poll sucks.)