I think d20 Modern was fairly succesful, but probably never could compare to D&D (but as others pointed out - what can?).
Unfortunately, Project Javelin and further adventure campaigns were never realized, and I think d20 Modern is "dead" now from a WotC support perspective.
I am not sure I would want to go back to d20 Modern now, with 4E around and many d20 Modern or OGL offshots of d20 Modern that probably show how much potential for a modern d20 game is still unrealized in d20 Modern.
I'd see as its failures:
- No real adventure and campaign support. The system is a toolbox, which is great in some ways, but it really hurts its ability to be used out-of-the-box and to motivate people to play it for a longer time. (I think that's a problem of many OGL games that consist mostly out of core books)
- There is not a lot of mechanical innovations for the non-combat stuff. Seriously, d20 Modern needs more material for creating organizations, running investigations and similar stuff. Skill Challenges or similar mechanics should have been made core already in d20 Modern!
- The talents are either boring (Oh, wow, Melee Smash lets me deal more damage And I can take it 3 times!) or needlessly convoluted (okay, I roll a Charisma Check against a DC to activate this power, and then the opponent makes a Save, and then he might get a penalty?).
Star Wars Saga did this a lot better.
- Toolbox not thought to the end. D20 Future was very interesting, but it also showed a few problems of the entire approach. Higher PL gear was plain better (well, who would have thought!), but there was no clear way to model this for stuff like CR or overall power level. Games like Mutants & Masterminds (Campaign Power Level) or Torg (Axioms, Value Chart) do this a lot better. The game was supposed to cover all this, but it couldn't really do it so well.
It also had a few good ideas (beyond the solid foundations of the d20 System):
- Wealth System. Really, you don't want to deal with Dollars, Euros and Yen in a modern game with its credit cards, banking account, speculation and rents. Abstract it. The specific implementation might have its flaws, I think, but the basic idea was right.
- The Multiclassing focus of the system. I would change the way classes worked, but the general idea of mixing and matching abilities of different classes or roles is sound. It avoids the complexities and imbalances of a fully fledged point buy build system, but still gives a lot of ways to tailor your character and create interesting archetypes (including unusual ones). For example, mix Strong with Smart!