Although 16S ribosomal RNA (rRNA) gene sequencing is well established for correctly identifying bacteria, its most efficient use in a routine clinical laboratory is not clear. We devised and evaluated a strategy to select gram-negative rods and coccobacilli (GNRCB) for which sequencing might be necessary before routine identification methods had been exhausted. The prospectively applied selection criteria were primarily based on the isolate's display of unusual or discordant phenotypic results and/or disease correlation. By using this strategy, we selected a total of 120 GNRCB (representing only ∼2% of all identified). The strategy was demonstrated to be efficient because the preliminary phenotypic identification for 79.2% of those isolates needed revision (18.2% were novel and about a third would have required further extensive testing). The knowledge that 1.6% (ie, 79% of 2%) of isolated GNRCB might benefit from sequence identification could provide guidelines for routine clinical laboratories toward efficient use of sequence analysis.