The simple answer is because they are not equipped to do so.
WOTC is not equipped with someone with the knowhow to hire a firm to properly produce a digital product?
That's not as straight-forward as it sounds. I've seen a number of cases where large IT departments (i.e. places you'd expect to find expertise) had a hard time learning new tricks. This includes the ability to correctly manage third-party developers.
The truth is, managing software development is hard. It's made harder by the fact that "managing" is largely a non-technical function, but actually doing is very technical. The paradigms involved have changed dramatically since many folks who "used to do IT" were actively engaged. Heck, there are a lot of really good developers who are just one cycle out of date. It's hard for the pseudo-technical and non-technical managers (literal or figurative) to sort out who's who. While you may be able to find a guaranteed top-tier talent, it's going to cost you. Odds are that someone in the food chain is going to try to keep costs down and you'll end up with a) talent with an attitude issue, b) folks who are competent, but don't quite have all their 'I's dotted, or c) an up-and-comer with all the raw talent to pull it off, but in need of a bit of discipline. If you're really lucky, you find someone with talent and a passion for the subject matter, but those folks are more likely to do a side project of their own (HeroForge, etc.).
Based on other projects, I could easily see the development effort alone costing a half-million dollars or more. A project I was on that had a project manager, business analyst/QA, and for developers ate through $750,000 in one year -- and we didn't have to include infrastructure costs because we were corporate. A team for DDI could reasonably be that size, given what people want for it (and what was promised). That doesn't even count the ongoing maintenance or digital content (
Dragon and
Dungeon). Also, that team would have fit into category "B", above (good for the task they were hired to do, but not for "shrink-wrapped" software).
There's an adage that goes, "cheap, fast, quality; pick two". Almost every project I've been on has tried to fix cost and time. Commercial products that lack quality never make it to market or are quickly discontinued. IME, cost is a function of quality and you can pay now or pay later. It should really be, "quality, fast, maintainable; pick two", but I digress.
The point is, doing DDI (or whatever name) right would not be cheap. I don't know how big the market is for DDI, but it may not actually support the software at a level sufficient to justify the investment -- or the powers that be may not have faith that the market is large enough.