Is one in five such a bad ratio for useful innovation?
It might speak to there being some problems with state-side testing of devices for appropriateness on the battlefield but a 20% success rate of cutting edge equipment seems pretty good. Infantry folk are probably going to use the useful bits and pack away the useless ones - that feedback will eventually trickle back to fulfillment and the kit will be updated.
Also - I have a lot of faith in infantrymen finding really creative uses for tech that folks in the lab might disregard.