My extended response is because this is something that I see very often.
The blame is put on "the designers" (in the abstract), but it is deeper than that.
> We found that we could not get a single entity to deliver a comprehensive solution.
I can relate to that.
But the problem is more than just the designers. It's a mixture of things. A comprehensive solution should include a comprehensive approach. That's the hard part.
It's hard to prioritize as a designer when user experience is not a priority. Take, for example, "undo". Every usability guideline will recommend you have an undo (a "trash" area in an email UI is an undo, too).
But having that functionality is an architectural decision (like most UX decisions). You can design that without issues if the whole team is on board. But when "design" is merely the "V" in MVC terms, there is a strong push for showing a visual confirmation: easier to develop and design.
When an agency does the design, you also have a monetary incentive: it's better for the agency to hand off the nice-looking design rather than enter into a negotiation with the engineering team about why it's essential to expend a month having the architecture to support that "undo" button.
Also, the "undo" button should be paired with principles applied across the product. You spent a month adding undo to a single part of the product, but another part doesn't have it, and instead of doing things better, it makes it worse.
This piecemeal growth is also common in engineering; the difference is that most users don't see that mess.
I wish to have a simple solution, but there isn't one. It requires a lot of work, from business strategy to process organization.
I apologize for the apparent generalization, and appreciate the reply.
However, it has been quite common.
Back in the day, my company worked with Pentagram for visual design, and DDO, for interaction.
We found that we could not get a single entity to deliver a comprehensive solution.
But, as a bit of a riposte, I did say "It has been my experience." It wasn't a generalization. It was a report of 30 years' worth of experience.