Milk, fat, and protein loss due to a new subclinical mastitis case may be economically important, and the objective of this study was to estimate this loss. The loss was estimated based on test-day (TD) cow records collected over a 1-yr period from 400 randomly selected Dutch dairy herds. After exclusion of records from cows with clinical mastitis, the data set comprised 251,647 TD records from 43,462 lactations of 39,512 cows. The analysis was carried out using a random regression test-day modeling approach that predicts the cow production at each TD based on the actual production at all previous TD. The definition of new subclinical mastitis was based on the literature and assumed a new subclinical case if somatic cell count (SCC) was >100,000 cells/mL after a TD with SCC >50,000 cells/mL. A second data set was created by applying an adjustment to correct low SCC for the dilution effect when determining if the previous test-day SCC was >50,000 cells/mL. Thereafter, the loss was estimated for records with SCC >100,000 cells/mL. The production (milk, fat, or protein) losses were modeled as the difference between the actual and predicted production (milk, fat, or protein) at the TD of new subclinical mastitis, for 4,382 cow records, and 2,545 cow records after dilution correction. Primiparous cows were predicted to lose 0.31 (0.25-0.37) and 0.28 (0.20-0.35) kg of milk/d at an SCC of 200,000 cells/mL, for unadjusted and adjusted low SCC, respectively. For the same SCC increase, multiparous cows were predicted to lose 0.58 (0.54-0.62) and 0.50 (0.44-0.56) kg of milk/d, respectively. Moreover, it was found that the greater the SCC increase above 100,000 cells/mL, the greater the production losses. The estimated production losses were more precise than previously reported estimates.