Sequencing saturation is calculated as:
1 - (# of unique molecules detected)/(# of total reads)
So for example, if you detected 800,000 unique molecules with 1M reads you would get:
1 - 800,000/1,000,000 = 0.2 saturation
If the sequencing saturation in your experiment is low, this means that additional sequencing will yield many more detected transcripts and genes per cell. As your sequencing saturation gets high (>0.5), you are approaching the point of diminishing returns and additional sequencing will be less productive in terms of detecting additional transcripts and genes per cell.