A few observations:
* I think it's a grave mistake to attribute the downswing in d20 sales with 3.5. The information I've had access to hasn't shown a slide downward that corellated with that event.
Instead, WotC's switch to full color interiors and hard covers, along with the institution of the development team, made it more difficult for d20 companies to produce supplements that could compete with WotC.
By 2003, the adventure market had already dried up save for those few companies that specialize in them. Thus, d20 companies already leaned on supplements, and the changes at WotC hurt them. The gap between a d20 book and a WotC book became cavernous.
In other words, it's a lot harder to see a big difference between Sword & Fist and a book like Quintessential Fighter. Aside from the D&D logo on S&F's cover, the books have the same basic level of production values. But if you put QF next to Complete Fighter, the difference is obvious and enormous.
Furthermore, the introduction of the development team adds a level of refinement and error checking to WotC's RPG department that no other publisher can afford. The development process is by no means perfect, but it's a step forward. As the team gets better and our processes improve, WotC's advantage here will only widen.
* The interesting thing is, if you look at the print market today, everyone is doing licensed games, new games, or supplements for those titles. There really aren't many companies still doing books that you can add into an existing D&D campaign. At the very least, the raw volume of stuff has slowed tremendously. Most of what comes out now is d20-based games or licensed material.
* With all this in mind, most publishers have taken a beating over the past year.
IME, the d20 companies made a number of mistakes since 2000:
* Nobody established a coherent identity early on. In the beginning, the d20 companies were just companies that made D&D compatible adventures. Privateer did the best job of avoiding this, as the IK had a unique, vivid look and a clear identity.
* d20 companies imitate rather than innovate. We had the shift from adventures to supplements. We had the shift from original games to the flood of licenses. Malhavoc dodged this by publishing variant PHBs, new games that were also easy to loot for D&D rules expansions/improvements.
* These shifts pushed d20 companies further and further away from what was supposed to be the core strength of the license - the ability to produce material compatible with D&D. Goodman Games still does well with its DCC line, plus it has the advantage of doing really clever things with the nostalgia angle that help it stand out.
* d20 companies have never been able to work together. The simmering angers, jealousies, and petty rivalries help sour what could be useful and profitable business relationships. We saw the beginnings of this in WW's Sword and Sorcery imprint, and I think that provided a good model for how companies can leverage their individual strengths to produce a stronger whole.
* Finally, and this ties into the point above, none of the print companies understand or utilize the open source nature of the OGL and the d20 SRD. We haven't seen any steady, incremental improvements in the engine. Everyone just reinvents everything all the time. I think the PDF market has the print companies beat here.