This web tool and associated paper by Prykhozhij et al was interesting as it may be very useful for certain target identification situations, particularly for highly similar gene paralogs and internally duplicated exons within genes.
If you'd like to target a conserved domain in a gene, you'll probably want to make sure the CRISPR guide RNA won't recognize the DNA encoding the same domain in other conserved members of the gene family. Conversely, there might even be special situations where you would want to zap more than one target at the same time by actively choosing to use a shared CRISPR target that's present across conserved genes. As the title implies, this web tool can find the shared and unshared CRISPR targets among a group of similar sequences.
I've also thought about this idea for picking targets in which a SNP would disrupt guide RNA binding. For example, in a situation where you'd like to bias in favor of specifically only modifying one of the two alleles in a cell, you could use a SNP-specific CRISPR to target a heterozygous embryo, cell line, or whatever, to increase the odds that only one of the two alleles is modified. (It's not a perfect idea as single base changes don't always change CRISPR target specificity - though they seem to usually do this if they are in the "seed" region.) This effect might be good in situations where you really want to generate a mutant allele, but the homozygous mutant is absolutely lethal. Mouse pronuclear injections are often done on hybrid strain backgrounds so there's lots of heterozygosity to be exploited for this. If it's useful. Maybe??
PLoS One. 2015 Mar 5;10(3):e0119372. doi: 10.1371/journal.pone.0119372. eCollection 2015.CRISPR MultiTargeter: A Web Tool to Find Common and Unique CRISPR Single Guide RNA Targets in a Set of Similar Sequences.Prykhozhij SV, Rajan V, Gaston D, Berman JN.