Have you ever picked up a new skill, or learned some new technique, and become completely infatuated by it, to the point where you're just itching to try it out? I don't know if there's a name for it already, but let's call it "Honeymoon Syndrome".
A case study
To give an example, as I started my first real job I had read a lot about Test-driven Development (TDD). And so, when I was given the honor to develop a new internal service, I was eager to try out all the things I had learned. The only problem was that I had misunderstood a few critical parts of TDD. I began by writing a test suite for the entire system, using stub implementations of all the interfaces I had planned out. I had missed the subtle detail of TDD that said you should start by writing a single test, make sure it fails and then make it pass. I had also combined this mistake with the "Big, Up-Front Design" anti-pattern. The end result was that I had to scrap the interfaces I had written, plus all the tests, and start over.
The problem
The problem with Honeymoon Syndrome is the convergence of maximum enthusiasm and minimum experience. Since the technique is novel, we haven't had time to figure out its potential drawbacks and pitfalls, which means we risk using it in ways or places that are not appropriate. This is fine if it's in the context of a hobby project or something else that's not time-critical, but can be catastrophic in the context of a deliverable with a deadline.
Dealing with Honeymoon Syndrome
How do you deal with Honeymoon Syndrome? I suggest the following rule of thumb: If you've never a technique before, don't use a critical project to experiment with it.
Another idea is to consult with a colleague who is not as infatuated with the new technique as you are, and explore the idea with them. If they also think it sounds like a good idea, make a time-boxed experiment and solicit feedback from other colleagues as you go. I've noticed a tendency for developers to disappear on "honeymoons" when they fall in love with a new technique, which means there's no one to talk them down if they take it too far—having a colleague chaperoning them makes sure nothing untoward happens.
In closing, I don't want to discourage experimentation in the workplace—after all, that's how you hone your craft and discover better solutions. However, remember that it is an experiment, and that—like all experiments—it might fail.
