Why sell a product that forces the user to build their rocket in an unconventional method?
Mainly because there's no other choice. The limitation here is physics, unfortunately. I'm sure Eggtimer Rocketry (and many others) would be happy to break the laws of physics if only they could. Antennas generally don't like being placed close to large pieces of metal, it kinda cramps their style.
Similarly things that transmit RF don't like being contained in tubes that absorb RF waves. There is always the option of using a remote antenna to position it away from those obstructions, as mentioned in the website/instructions and I believe on this thread. I've also seen several designs where people point the attached antenna outside the top or bottom of the avionics bay which gets it clear of the metal, though I suspect it might see a lot of abuse there during deployment/landing also.
Eggtimer Rocketry does have separate products for deployment and tracking, which work well if your rocket has the space to mount two separate pieces of electronics, you can put the Eggtimer (deployment altimeter) in the avionics bay and the Eggfinder (tracker) in the nose and provided the nosecone doesn't absorb/distort RF that will work just fine. If you don't have the space, there is the option of the Eggtimer TRS, which combines those two functions into one. But it does bring with it the complication of either doing deployment from your NC (which, if you're limited on space as I am in one of my rockets, may be the only choice anyhow, likely better suited for a cable-cutter-style approach than the more traditional drogue+main) or if you're going to place it in an avionics bay then there is the issue of having the antenna near other pieces of metal, so you either have to just deal with that and the reduced range / dead spots that result, or else go with a separate antenna that can be mounted farther away from the metal. If there was a suitable non-metal solution for an avionics bay that would be a third possible option, which is what the OP inquired about. But the same would be true of any other RF transmitter solution, at least in the 900MHz band, but I don't see why it wouldn't affect other bands as well. Even the GPS receiver antenna won't much like being blocked by metal if it's in the wrong place.
About the only other option would be to just not fly an RF transmitter at all. That's not acceptable/desirable to everybody of course. I launch in fields where without some sort of tracking system I'd have lost all of my rockets already, so I won't fly in those fields without a tracker. My Eggfinder paid for itself the very first time it flew, even though I thought I saw exactly where the rocket landed.
To the OP, you didn't really mention how big/heavy your rocket is. Are you looking at the TRS because it's a low-cost solution, or because you have a smaller rocket and don't have space for two separate units? If you have the space, separate units is something to consider (the Eggtimer is more capable than the Eggtimer TRS from what I've seen, more channels, more options for the channels, etc). I haven't used it myself, I have some Eggfinders because it was a low-cost, no-license-required tracking method for rockets where I already had (or otherwise didn't need) deployment electronics, and a TRS for a rocket with no space for separate solutions where I fly it in the nosecone, so nearby metal isn't an issue.
One thing I'd wonder about, having just looked at a fully 3D-printed rocket this past weekend, is whether there could be some sort of (probably pretty thick and heavy) plastic avionics bay where two coupler-style pieces could thread together instead of the more traditional tubes, end-caps, and all-thread/nuts to hold the end caps together. I'm sure this would only be good for lighter rockets, at least without being a significant hunk of plastic. But I would imagine such a thing could work in some circumstances at least, in the case of these 3D printed rockets threaded sections were what held the entire rocket together (since the printer couldn't make a single, long airframe). The ideas of passing the shock cord through the bay also sound interesting (tackle the problem of stresses being placed on the end caps instead of an alternate way to attach those caps), though there would still be the question of how you'd hold the bay shut I'd think. Perhaps that's with more nylon/kevlar? Making a good seal that way while still allowing easy access to the inside of the bay when needed sounds tricky to me. Though there was
the roll-pin idea mentioned on this thread, I wonder if something like that could be applied here.