Organizing for Product Design and Production

I had an interesting chat yesterday with a manager in a large international company that is restructuring to support better innovative product creation. He asked my opinion on a number of topics, relevant to their effort:

• Matrix organization of cross-disciplinary teams: marketing, engineering, and operations, for instance. I suggested that at the innovation phase before product development, the team might look different and have different skills, such as design!
• Does long-distance collaboration really work during product design? I said it can, and that although video gets a bad rap, we used it effectively at Adobe in the same time zone. I am less sanguine about major timezone differences, and still a firm believer in regular face-to-face, especially at the start and during times of tension, for defusing it over a drink or meal. For designers, there's more artifact production to communicate ideas long-distance; and the whiteboard collaboration hasn't really been adequately replaced by any conference tools I've seen, but I'd love to be wrong if someone has a suggestion.
• Quality of product: Get it out or get it right? I said this depends on ability to absorb risk in a brand, and first to market vs. competitor situation. A first-to-market product can get a few things wrong, and still have the advantage. Not if they got the major things wrong, though, and destroyed their own business and can't absorb the hit. Soft launches and long betas are another tried and true method. Google seems to live on eternal betas -- but of course they did get search pretty much right, first, as far as state of that art goes now.
• Outsource concept design or do it inhouse: I said this depends on how much product innovation is meant to be a core competency. I think it should live in-house if it's meant to be core, because it's a risk to outsource all good idea creation. Also expensive.

It wasn't a specific question, but I pushed for the value of ethnograhic research for identifying new product areas and gaps in markets; and the value of design of both the industrial variety (closer and closer to home for software now) and the usability/interaction variety. The need to reward and remember failures is important here to encourage iteration and re-application of old ideas in new ways.