If the gRNA works, chances are there is a small indel at the cutting site, but how frequent will this happen? Can I use the length difference as an indicator of gRNA efficiency.
the chances just depend if the plasmid transfection/nucleofection of the RNP were successful. If the gRNA is properly designed (PAM sequence, uniqueness, etc.) it will always cut. Repair after NHEJ is totally random, and usually about 1 -10 bp are deleted but insertions are not rare either. Therefore, the length in difference cannot be an indicator for gRNA efficiency as, as long there has been been some indel and you verify it by sequencing that the target protein is disrupted, all gRNA are equally efficient.
I let you this interesting Addgene blogsite: http://blog.addgene.org/crispr-101-non-homologous-end-joining