I'm done with the phrase "experimental poetry".
Starting with the literal, let's combine the definitions of experimental and experiment to define the term (all definitions from the Random House dictionary via Dictionary.com):
"pertaining to, derived from, or founded on a test, trial, or tentative procedure"
No. Poetry isn't tentative. More on test and trial in a minute.
"pertaining to, derived from, or founded on an act or operation for the purpose of discovering something unknown or of testing a principle, supposition, etc."
Tempting, because we do write to discover what's on our mind, but to term this experimental would be in ignorance of the difference between experimentation and exploration. When no response is measured ("If I do this, what will this thing do?"), discovery is an act of exploration - a documentation of what's already there (though perhaps unknown to you). Excepting acts of improvisation with a live audience, there is nothing to measure in response to the act of putting words on the page or screen.
"pertaining to, derived from, or founded on a chemical experiment OR a teaching experiment OR an experiment in living."
You might pass muster applying this definition IF the rubric or theme of the poem fits one of these definitions. However, there's nothing chemical and nothing you can live in when you're presenting a poem. Surely you can create poem as a teaching example, but that's no more an experiment than the varnished plates of food outside the Greek restaurant in the Food Court at the Mall.
"of the nature of an experiment; tentative, as in "the new program is still in an experimental stage".
Boo on tentative, but the rest has potential. However, it falls short when you realize that this means "having the potential to fail" - if the program tanks and is abandoned, it was a failed experiment - OK. But if a poem fails and is abandoned, we have a word for that: a "draft". I contend that nothing that reaches your personal definition of complete can be termed experimental.
"based on or derived from experience; empirical"
This covers everything you'll ever write.
A quick word on test and trial: a test is an action you take or a situation you create for the purpose of seeing what will happen. But it's not an experiment until set it leaves your control for the trial whose results you will observe objectively. If the piece of art you create is never outside your control, it can never pass this definition. This leaves in play the performance elements of some poetic presentations, but unless reactions are noted and changes to the poem considered, this too fails the definition.
Am I being deliberately too literal to make my point? Let's consider what turns up on the first page of a Google search for "experimental poetry":
Text Etc. discusses many different elements of poetry. About "experimental poetry", into which bucket it collects VizPo, Concrete, Conceptualism, and Code as Text, it says: "Experimental poetry can be intriguing and pleasing, but it is not poetry as commonly understood by the term, and has therefore to be judged on different grounds, most commonly those of the visual arts, which it increasingly resembles.". OK, no argument here. But if it's "not poetry", then why try to make it so - let it be a new art form, not a "poem". This is how science and engineering operate, isn't it; when the experiments begin to produce a new and unique set of rules about the systems they define, they get classed out as a new discipline with the old as its root. For further consideration, Goef Huth, who wrote on visual poetry in Poetry last year, starts his bio with "Geof Huth is a poet and visual poet..."; why do we need for the distinction if one is merely a special case of the other? Bottom line: we don't*.
In her essay at CORNER, Laura López Fernández notes that "Contemporary experimental poetry, in its various forms of manifestation -- visual poetry, phonetic poetry, sound poetry, performance poetry, non-object poetry or action art, video poetry, cyber poetry, computer holopoetry, mail-art, etc. is an integrative and interactive art that requires a "reader" willing to participate in a new configuration of semiotic codes." OK, if you know that semiotics is "the study of signs and symbols as elements of communicative behavior", and you are deeply enough integrated with contemporary poetry to accept poetry without symbols (metaphors, allegory, etc.) isn't poetry at all, what is experimental here? It's just the integration of technology or performance. These are neither new nor non-obvious extensions of poetry. BUT... if it really does require a new science, then it's a new discipline and a new art. See above.
Nothing else on the front page of the Google search attempts to define the term, though Selby's List provides many paths to explore it. There's one very silly definition that actually lists a dozen or so forms to apply if you want to create an "experimental" poem. This is the most egregious misapplication of the term and does disservice to the serious definition attempts above. Make a note: if a template exists, you're not conducting an experiment.
Let me now air the biggest flaw in this discussion: I've turned to Google, which returns results based on popularity, which isn't fair to any art form clearly operating on an extreme edge of its accepted definition and therefore not likely to be popular. I get that. But point me at a credible source that more completely explores the test-and-response definition essential to experimentation, and we can talk about the rest.
This is another area where being a career technologist disposes me to a particular interpretation - and one that in its own way, is a "fringe" opinion for someone who studies the art of poetry. I get that, too. Teach me where I'm wrong.
A final disclaimer: Don't take this to mean I don't enjoy some of the work currently showing up as "experimental". I'm really just becoming confident enough in my understanding to begin seriously tackling the copy of Poetry for the Millennium I got as a gift years ago, but I'm not ignorant of of the works of Silliman, Armantrout, Rothenberg, etc. That's not the point of this at all. The point is this:
Don't marginalize forms or create barriers that prevent new audiences from encountering poetry by branding it with a scary and inappropriate term. Just make it, and just take it in.
*BTW, You should go spend an hour on Geof's website; just because I disagree with the term doesn't mean it's not well worth your time.