Agree completely with the agency problem. There are far too many folk who confuse the agency model of doing design with the way design is done. It's one of the reasons that the Lean UX community of practice get some bad press in parts of the design community.
(Although the agency problem isn't really lack of iteration. Pretty much every design process is highly iterative. It's just that the agency folk do it by themselves and, usually, up front.)
I find the second point much less convincing. I know lots of folk who love doing long-term product work. I guess I see about the same short/long term work preferences as I do in the development community. You get to be a good designer by doing lots of design. As long as the in-house opportunity has that, you can attract designers if you're recruiting the right way.
From your last paragraph it sounds like you outsource most of the visual design work, and in-house things like interaction design, generative user research, user testing, etc.
That works well for some products (I'm guessing that what you're building is a B2B or B2C web or desktop app that uses a "normal" UI without unique interaction models, and with a market that isn't design focussed - but I could, of course, be way wrong there).
Other kinds of products really need in-house visual design skills - because not having them gets in the way of that rapid iteration that you so correctly identify as being essential (games, some kinds of mobile apps, products with custom interaction models, markets that are design focussed, etc.)
I agree that the agency design process is internally iterative. But they don't ship or observe actual usage, so I can't call it real iteration, which I think requires a closed feedback loop.
I agree that there are a lot of people who love doing long-term product work. But I think to be amazing at a particular sort of design, you have to do a substantial variety of it. I don't think most startups offer anything close to the variety that an agency or a freelance career can provide. Which is why we turned to external specialists for things like our logo.
I agree entirely that about different contexts needing different skills in house.
Hi William! (guess who didn't look too closely at your username before :-)
Well, depending on the agency, there is a closed loop with the user - it's just a pre-product closed loop built around interviews and prototypes. But I generally agree that the agency model sucks quite badly for product development in startups (and a whole bunch of other things too come to that.)
I don't believe that's a closed loop. The purpose of commercial products is to sustainably deliver value to purchasers. Until that is achieved, the feedback loop is open.
There could be some design firms that actually go that far, but I've never heard of a project like that. I do agree it's better when design shops attempt to test their design hypotheses through prototypes and user testing, as that does leaven the dreams with a dose of reality. We do a bunch of that for exactly that reason. But we never mistake a successful user test for actual success.
Obviously product based feedback is "better" on some levels, but it can get in the way of learning the answer to some questions quickly I'd much rather just get the answer quickly ;-)
As I play with experiment/hypothesis model I'm more convinced that increments-of-learning rather than increments-of-product help me more - at the early stages anyway. Since many of the folk I talk to seem to be really bad at validating business questions via early customer contact I think there's an opportunity for external folk to come in and help.
I'm really interested to see what agencies like Proof (http://proof-nyc.com/) organise themselves and their work with clients. I know my relationship with some clients is more like design/UX coaching rather than more traditional agency type work.
(Although the agency problem isn't really lack of iteration. Pretty much every design process is highly iterative. It's just that the agency folk do it by themselves and, usually, up front.)
I find the second point much less convincing. I know lots of folk who love doing long-term product work. I guess I see about the same short/long term work preferences as I do in the development community. You get to be a good designer by doing lots of design. As long as the in-house opportunity has that, you can attract designers if you're recruiting the right way.
From your last paragraph it sounds like you outsource most of the visual design work, and in-house things like interaction design, generative user research, user testing, etc.
That works well for some products (I'm guessing that what you're building is a B2B or B2C web or desktop app that uses a "normal" UI without unique interaction models, and with a market that isn't design focussed - but I could, of course, be way wrong there).
Other kinds of products really need in-house visual design skills - because not having them gets in the way of that rapid iteration that you so correctly identify as being essential (games, some kinds of mobile apps, products with custom interaction models, markets that are design focussed, etc.)