Split decision: A/B testing online is a nice article introducing a topic some people may not be familiar with. The idea is you test one version against another to see which performs better on some benchmark.
5-Second Tests: Measuring Your Site's Content Pages is even simpler. "It has a similar structure to traditional usability tests—you'll have users, tasks, and the site you're testing. The difference is in the specific protocol for running the test."
Testing works. Asking for opinions, on the other hand -- not so much.
First Rule of Usability? Don't Listen to Users explains use is different from talking about use. Self-reported data is three layers of abstraction removed from reality.
Those are some good points.
I was just reading about those exact things in a book which I grabbed from the library today:
designing the obvious
a common sense approach to web application design
by Robert Hoekman, jr. - published 2007
It mainly talks about web application usability and design but the same principles can be applied to web design in general. One of the main points is about only designing what is absolutley neccisary, and forgetting those 'cool features' for your initial design. File them away and look at them at a later date and evaluate once your site or app has been 'live' for a while, whether or not you still want/need them.
My favourite point of view from Hoekman is 'Start ignoring the demands of users and stick to a vision (gasp!)' - so many times ive tried to design something but get caught up with trying to accomidate every single one of the users views or opinions that I lose what my original intention was.
The book is a good read on web design in general and I highly reccomend it.
User triangulation: how to listen to customers is great for product development, but also for writers.
While it's okay to ask for some critique in a forum, you really have to be careful. Yes, you'll get some opinion, and a few code or spelling corrections which can be helpful. But few people will actually evaluate a site like a user or customer would -- despite every protestation to the contrary.
What you won't get is anything which will substantially improve the product, software, website, or writing. Yet how many people will slap up a web based survey -- Cooper calles it the customer-driven death spiral -- and others call surveys deathwish research.
I call them the "hit counter" of the CMS era. They're a favorite toy for content management systems, but surveys can be tricky ...just ask President Kerry. And there are better ways to get actionable feedback you could code into a CMS module. But yet again, people ask what they can do with technology and don't study how to design systems for actionable feedback for the user.
How many people will read this thread, nod their head, then post a site in the forum to get a casual look-see opinion from someone who is completely unlike their target user. At the very least, if you wanted feedback, you would post a user profile. A profile which makes it clear how different the target user is, their technical background, their motivations and tasks.
A related thread you can check out is using personas for writing.