I've used grey cards hundreds of times with my spot-meter to identify and fix a mid grey value in complex subject matter. But it has been a frustrating process easily subject to error.

The problem comes from the fact that grey cards are not perfect Lambertian reflectors. The"greyness" varies with the angle of the card versus the angle of the light versus the angle of the meter. In practice I'd select a meter reading when the card showed minimal glare and no shadow; concientious guesswork in effect. A good incident light meter is better.