There seems to be a lot of unquestioned arbitrary rules inserted into place for what people think a trigger is in hypnosis.
The actual definition is very simple, at least from my perspective. A hypnotic trigger is anything that causes recall of a hypnotic suggestion. Anything.
The trigger itself doesn't need to be reinforced or have even been said before for it to be a trigger so long as it causes recall, it does not have to be a word or a sound. So long as the subject's brain can recognize it as a distinct thing, then it counts. It's closer to ontology than anything.
That being said, you need to understand properties of triggers and how every single aspect of a trigger will affect how the conditioned suggestion is received.
One of the most common and baffling mistakes I see both in trances and in triggers is the "active time" of the trigger; that's to say, the amount of time the brain needs to be acting on that suggestion.
I've always despised visual of an elevator or an ice block, because the subject needs to be hyperfocused on a very slow climb and fall which is near impossible to maintain the consistency of, especially if they need to focus on other things too.
In contrast, a trigger or suggestion that takes place near-instantly or over a few seconds only requires a moment of making an active change, which is simpler and more consistent, and allows for variability in the strength of the effects since you don't have the last minute of change to compare it to.
If this drop trigger only worked 80% as well as the last one, it's a lot harder to notice with a chunk of time of nothing happening between them. Even consistent, small bursts are better than a constant change, like stepping down stairs instead of going down an elevator.
Snaps are the most common form of signifying a trigger because we can make it with our hands, but more importantly, because they're sudden and loud.
The aesthetic of a sudden and loud sound translates to the subject's brain how to interpret the feeling of that trigger. This also means, though, that if you have a trigger that isn't meant to be a burst, that you may not want to use a snap at all.
This also extends into the idea of "suggestion equivalency", as I would call it.
Essentially, you want the aesthetics and potency of the trigger to reasonably match up to the effects of the suggestion itself. The reason a snap is powerful is because you're condensing a very strong effect into one moment. However, if you try to stretch that intense potency to multiple seconds, it'll simply fail. There is only so much focus, there is only so much trigger potency.
One of the myriad reasons I despise key elman inductions is the suggestion of "doubling trance" every time you blink. There is a cartoonish, absurd potency required for that idea that anyone who knows hypnosis at all from the hypnotist or subject side (or just like, the wheat and chessboard tale) would get severe cognitive dissonance from attempting to follow it.
Another issue to consider is the generality and widespreaded-ness of a conditional trigger.
One file from MissLilith that I think was rather silly (and I'm only ragging on her here because I like her work) tries to give the suggestion that any time you see any plant at all, you get aroused. Furthering the idea of keeping up intensity over time, if a stimuli is so ubiquitous that someone experiences it constantly, the spectacle of it is simply gone. That suggestion will, optimistically, exhaust itself within hours.
That's not even to mention that fact that seeing something does not have nearly as much of a spectacle and intense impact as a sudden sound or word. There's a reason why movies have stingers to signal sudden scary moments or realizations.
From a community perspective, you should also avoid the most common trigger words unless you want someone to be dropped accidentally or have your suggestions lumped in with ones from other hypnotists and files.
The last issue is another practical one, the idea of stimulus generalization.
Shortly after an intense session I had involving snaps one time, I watched a 2 hour long video by Nexpo about Petscop (good and atmospheric overview, though Nexpo is terrible when it comes to actually analyzing media himself).
Thing is, the text changing sounds in the game sound a scary amount like a snap. So, every single time there was gameplay audio, I'd get triggered by it, essentially fractionating myself due to continually bringing myself up right after it happened.
Clicker sounds can be set off by pens, snaps by other sudden, high-pitched single noises, even general obedience suggestions can be triggered by another person to your subject if they act similar or have a similar sounding vocal tone to you, among other things. One must always consider that possibility.
Overall, words as triggers are common for a reason, but you can do essentially anything, so long as you recognize many of the common shortfalls of doing so.
Good luck, and happy hypnotizing.