i've used non-hardening modeling clay and it never moves, even in direct sunlight for hours on end in 105 deg F. days (my apillar pods have some).
when used on doors they are usually deadened around anyway so it can't go anywhere. the purpose is to reduce transfer of basket vibration to the door metal. and it does work well for that.
i do strongly argue that we need sound absorption in the doors, but the only way to achieve it isn't with foams but good 'ol fiberglass batt insulation. since doors get wet when it rains you need to prevent it from getting wet in the door cavity, so i put it in 1mil PVC bags and secure it to the outer door skin with aluminum foil tape (sticks to metal very, very well). the result is the absorption we desperately need in the door cavity. without absorption the rear wave bounces around that very reverberant chamber and comes back out the cone later as phase interference. it's why simple songs can sound good but complex songs become "muddy". my approach works well and has resulted in some 1st place finishes in several organized comps. for comps you just have to be careful what class you run as in some it is against the rules to put stuff in your door.
when figuring out how much to put in the door, the answer is simply as much as possible on the outer door skin. the location of the door means you will achieve good midrange and midbass absorption as glancing angles increase the perceived thickness of the material. absorption starts at 1/10th wavelength thickness and is excellent at 1/4 wavelength thickness. the length of the door (and i have bags that run the entire length in three rows) will give good broadband absorption. the only trick is making sure it doesn't interfere with the door mechanism. this is much easier to do BEFORE you seal the door, but seals on the door should still be removable.
along with door treatment is the need for seals. we need to isolate the rear wave from the front otherwise we get cancellation. in a stock door, the plastic door panel is the front baffle of the speaker enclosure. weatherstripping foam is cheap and the most effective seal you can apply (closed cell). seal both sides of the basket - front and rear. getting the sponge foam rubber weatherstripping can also take the place of non-hardening modeling clay for basket isolation from the door/baffle. i also close up all holes in the inner metal door panel. i use a variety of materials but all are removable. my build logs show all of this effort in several different methods.
deadener has the purpose of reducing the resonance in metal. knock on the door and listen for how long the resulting sound lasts, or "rings". deadener should provide damping, and reduce the "ringing". the result is audible and simply knocking on the panel will give you an idea of it's resonance frequency and also the amount of damping applied. when it starts to sound like 2" of wood, you are reaching your goal.
to summarize, you are trying to make your crappy door made of thin metal and plastic to behave like a $10,000 home audio tower speaker that is acoustically dead in order to only hear what comes off the front of the cone and nothing else. that's the goal anyway. it doesn't take a lot of money to pull it off, you can get creative with materials. but it does take a lot of time. 10-20 hours per door is a good estimate. the effort is worth it and required to actually get the most out of the installed speakers.
cheers