Clinical exome sequencing (CES) has greatly improved the diagnostic process for individuals with suspected genetic disorders. However, the majority remains undiagnosed after CES. Although understanding potential reasons for this limited sensitivity is critical for improving the delivery of clinical genomics, research in this area has been limited. We first calculated the theoretical maximum sensitivity of CES by analyzing >100 families in whom a Mendelian phenotype is mapped to a single locus. We then tested the hypothesis that positional mapping can limit the search space and thereby facilitate variant interpretation by reanalyzing 33 families with "negative" CES and applying positional mapping. We found that >95% of families who map to a single locus harbored genic (as opposed to intergenic) variants that are potentially identifiable by CES. Our reanalysis of "negative" CES revealed likely causal variants in the majority (88%). Several of these solved cases have undergone negative whole-genome sequencing. The discrepancy between the theoretical maximum and the actual clinical sensitivity of CES is primarily in the variant filtration rather than the variant capture and sequencing phase. The solution to negative CES is not necessarily in expanding the coverage but rather in devising approaches that improve variant filtration. We suggest that positional mapping is one such approach.Genet Med advance online publication 06 October 2016.