In the book “Creating Motion Graphics” by Chris and Trish Meyer, Chapter 14 talks about the camera and they say it’s generally a good idea to animate your subject or animate your camera, but if possible, not both. This just helps you keep your sanity as the composition progresses and the movement gets more complex. The more values you have keyframed, the harder it’s gonna be to go back and altering things later.
When dealing with text, I try to move the camera as little as possible. The further you get away from it’s original default Position/Rotation/Orientation values, the harder it’ll be to control especially when you you find your position is 37 thousand pixel away from your starting point... . Typically, If I do find myself wandering “far from home”, I’ll often look for a way to PreCompose what I’m working on and transition into a new comp with a fresh camera.
Unless you’re really planning on doing 3d rotation or camera movement, there is no need to switch your text to 3d. Instead of using the text’s z position to have it zoom toward the camera, try just using scale instead... you’ll get the same motion blur and you might find it easier to control. Another reason I’d encourage you to leave things 2d when you’re able is because of the “Align” feature that isn’t an option with 3d layers. If you have 6 text layers with a “centered” justification, all you need to do is go down to the align window and with one click you’ll have perfectly centered words... either as an average of all the layers you have selected or centered to the composition as a whole. You’ll notice when you flip on 3d, this is no longer available to you because After Effects is looking for just X and Y values alone, throwing a Z value in their screws it up.
All that said, my basic point is keep things as simple as you can whenever possible... Now certain effects are going to entail some pretty thick and chunky timeline markup, but if you don’t have to complicate things.... don’t.... :)