My everyday targets are made from retroreflective tape, which works fine at 100 meters, but that's closer than I'd prefer, considering that SAR practice counts on seeing mirror flashes at 5 mile range.

One thing that is superb for a signal mirror practice target is an optical corner cube. At a range of 0.5 miles, the return from a 1.5" diameter corner cube is bright enough to "burn through" the fireball from a retroreflective aimer and let you know you hit the target. I wouldn't try signaling that cube at a range of less than 1/4 mile - too much like looking directly into the sun.

Without a "fireball" in the way (e.g., foresight or rearsight aiming) the return from the cube is visible at much longer range. I saw the return from the 2"x3" stainless steel mirror from a cube I'd mounted 3 miles away. I figure this approach drops off with at least the 4th power of distance, but even so, an array of 16 such cubes should work fine at 5 mile range (I tried it at 11.1 miles - no joy).

Alas, optical corner cubes are generally pricy, but occasionally you can get deals on Ebay. I've seen do-it-yourself plans for making a corner reflector from 3 mirror tiles, but I haven't tried that approach.

In a way, corner cubes and other retroreflective targets are too easy - after you hit them once, the feedback helps you hit them again. They are good for initial familiarization and drill, though.

For scoring, a spotter or video camera is better. At middling ranges (e.g., 1-10 miles) a star filter on the video camera makes it easier to analyze the video results later.

Two instructors and two groups, at range, is ideal.
For long range mirror signaling as part of On Target activity, we use cell phones and amateur radios for coordination.