While I believe that it works for you, your "early product design" process doesn't sound anything like interface design. It's just a subset of branding. Design is solving problems, not just choosing what the buttons on your website will look like before you have your product or any kind of context (the order of which sounds backwards to me.)
> And it’s not all about looks, you know.
You say this, and maybe I'm missing the point of your post, but you continue to describe how this process is entirely about how things look.
Additionally, I have an issue with this:
> Over time, you’ll build up a collection of “liked” work that you can show to any designer to give them a concrete idea of the style you like.
This sounds like a horrible idea. Any decent designer out there knows that Dribbble exists and doesn't need you to point that out to them. They likely have their own "style" (if we're reducing "design" to just aesthetics for the purpose of discussion) and won't like you showing them what "style" you want them to adopt. Their job as a designer is to solve problems, not make things look pretty in Photoshop according to your taste parameters. Everything they produce should adhere to your brand guidelines, but that's it.
I think the takeaway is that it lets developers be the designers. Most programmers I know are pretty good at solving problems, but they don't always have an eye for visual details aspect of it. The style guide gives them a framework to work in, and can hopefully prevent a UI atrocity.
Maybe hiring an outside designer to do the whole app would get you better results, but if you're going to toss all of that as the app changes, it might not be worth the expense. Assuming your devs aren't bad at user interaction stuff, this gets you most of the way there at a small fraction of the cost.
Articles like this one come up regularly, so let's not discount the method just because the word "design" doesn't mean anything.
"Design" can mean anything from Illustrator graphics to market research, interaction design, or even implementation. It's a dangerous trap to assume skill and experience in some of those areas equates to skill and experience in all of them.
The author of this article has clearly identified the intersection of need/acceptable cost, he just happens to be using an excessively generic term to describe the role. It's still a valid model, especially if the team already has talented product thinkers who are capable of following a style guide.
Yes, HN has ruined light grey on white for me as I tend to see it as a downvote. Seriously, I'm not being funny about that - I have begun to perceive low contrast light grey on white as something suspect.
A style guide is an important way to keep everybody on the same page. Not enough products have one. As others (including the OP) suggest, this is not the entirety of design, but it's a great step after something like Bootstrap to help differentiate the product.
Curious to get the hacker perspective on this topic...have you tried this method with your products? Tried something else that worked well? Let's chat :-)
Do you ask the designer to create the basic styles / example markup as well? General things like buttons, alerts, perhaps some generic layouts. Or just take the template they provide and have your in house devs hash that out?
Haha, sounds like me. Using bootstrap and playing around with the styles, I can get the project 80% of the way there visually. I never feel satisfied with my designs though... having a style guide to follow would probably be able to bump me past that.
We didn't choose that visual up front, we worked with a designer that chose that look based on the product (Minimal Analytics). The brand has defined the product over time and is very recognizable.
Whereas a flat css theme wouldn't have been unique or recognizable.
> And it’s not all about looks, you know.
You say this, and maybe I'm missing the point of your post, but you continue to describe how this process is entirely about how things look.
Additionally, I have an issue with this:
> Over time, you’ll build up a collection of “liked” work that you can show to any designer to give them a concrete idea of the style you like.
This sounds like a horrible idea. Any decent designer out there knows that Dribbble exists and doesn't need you to point that out to them. They likely have their own "style" (if we're reducing "design" to just aesthetics for the purpose of discussion) and won't like you showing them what "style" you want them to adopt. Their job as a designer is to solve problems, not make things look pretty in Photoshop according to your taste parameters. Everything they produce should adhere to your brand guidelines, but that's it.