The single most important factor affecting somatic cell count in milk is mammary gland infection status. In comparison, all other factors are minor. Consideration needs to be given to diurnal effects on Dairy Herd Improvement a.m.-p.m. sampling schemes. Somatic cell count linear score of 5 (283,000) appears to be a good choice of threshold for mastitis control applications. A greater understanding of the nonbacteriological factors affecting somatic cell count is needed so that relative thresholds could be used to improve the clarity of somatic cell count interpretation. Linear score loss estimates are effective educational tools providing motivation for mastitis control implementation. Infection status or milk loss estimates based on single somatic cell count tests on individual cows are weak. A lactational average linear score on individual cows or linear score compilations across a herd provide credible estimates. Treatment of subclinical mastitis based on somatic cell count levels is not economically beneficial and is not recommended. Usefulness of Dairy Herd Improvement somatic cell count data as a mastitis management tool requires measures of mastitis level, new infection rate, and mastitis pattern within the herd over time.