a few things i've noted for myself.

the first test ever (or maybe even the next season) is a shot in the dark. Pacing and effort get MUCH better over time and over iterations. You get better at 'testing' the more you test. I like to try to hold a flat line vs negative split, limited standing.

i can't find the picture at the moment to support, but for me I just do my good warmup , (whatever u think u do, lactate buffering, progression blah blah) and then 5 minute rest and then the 20 minute full blast. I have tested a decent amount to not negative split the effort, but from the final 20m average i take 93%.

The picture i was trying to find, implied to whatever degree that based on the type of type you might be and whatever physiological systems you got, u might want to take 93% vs 95%. I've used 95% in the past and found it a bit too hard. And interval repeat ability can provide the necessary insight for this. If your '100%' FTP number is 300w, and you cannot do 300w for 5 minutes in a workout, something is awry. What others have said isn't wrong either. You could in theory, not ever test, but do workouts and gauge 'fitness' from there. IMHO as long as whatever number you work with to set zones is do able in efforts and workouts, that's probably the best estimate to your FTP. It doesn't have to be exact to a decimal. I test about 3 times between dec-march. Then once mid summer and 2-3 times in the fall. It doesn't have to be every month or bi-weekly. If the majority of my workouts are indoors, i test indoors. If im riding outdoors, then i'll test outdoors, and i've personally found not that much variation between in vs outside tests. Some claim they magically find 30w outside...