Every time sites like Facebook or Google introduce something new, legions of people get emotional about it. All the angry fist pumping is causing a scenario where creators are scared to change things or make new things. Consider for a moment the financial impact around small changes to little things like Google search link colours, which can affect the link count.
When you consider the large numbers these companies are expected to make from click revenue, you can appreciate how difficult it must be change things, when one small mistake can lead to millions in lost revenue.
This is when you start to truly understand the value of testing things.
Don’t be afraid of change, don’t be scared to try new things, that colour change could equally increase revenues. If you have a new idea, try it, test it, if it succeeds, then brilliant but if it doesn’t, put it in the big pile of tested ideas. Who knows when users might see the value of that idea in the future.
Many great ideas are incarnations of earlier ideas, the thing about the best ideas is timing, and understanding user needs.
There are many avenues for reasonable testing, through wireframing, proto-typing and even testing buckets. Be sure you test your development against the kind of people who will use it, don’t assume they 'think' the same as you.
After years of doing this, the only thing I expect from the user is that they will do the unexpected, which is why testing is so important.
I’ve provided some relevant links below.