I usually use some combination of incident and multiple spot reflective readings to determine exposure placement and the correponding development plan. The exact method depends, I think, on the subject and the nature of the backlighting, plus whether supplemental lighting (fill) is possible and/or desireable. "Backlighting" includes a lot of variable situations.
For example, a backlit portrait or model shot would be handled much differently than an image of a translucent object where the light transmitted through the object might determine its luminance value. With the former, I start with a spot reading of the face, and adjust that to place skin tone values; then incident-meter the back light to see what that's going to do with hair, etc. If the difference is too great, I might add a reflector or fill flash to bring up the light level on the subject. For the latter, a spot reflective metering of the translucent area, adjusted to place its tonality might be given precedence.