You mean use fuel to remove all horizontal velocity? Yes, that works fine and puts much less stress on the exterior, but it's a gigantic amount of fuel, almost as much as it took to get into orbit in the first place.

How would you do that?

Keep in mind that the object orbiting is already falling. Orbiting earth is literally "falling around the earth", compared to "falling down to earth" which we are more familiar with from throwing rocks and whatnot.

So to go "straight down" either it would need to orbit the sun (instead of the earth) and have its orbit intersect that of the earth, like the meteors we're worried about, or it would need to do a very strong deceleration burn.

You still need to reduce your horizontal velocity. (For "geo-stationary", think "really high and really, really fast")

Either you do that with atmospheric drag, or a huge amount of fuel. The weight of heat protection is much lower and more efficient than the fuel option.

Coriolis force. If a geostationary satellite was instantaneously accelerated straight down towards the Earth, its path (from Earth's perspective) would not be straight down as its original circumferential velocity at geostationary orbit (pi * geostationary orbit diameter * Earth radial velocity) would be "too high" when the altitude is lower.

To continue moving "straight down" its angular velocity would need to be constant, which means as its altitude decreases the circumferential velocity would need to decrease. But gravity only pulls down, there's no force to accelerate it in that direction. So therefore it appears to curve off to the side.