The aims of this study were to estimate milk yield (MY) losses associated with subclinical intramammary infection (IMI) in dairy goats and to assess if somatic cell count (SCC) can be used to estimate such MY losses. We used 2 data sets to study these questions. The first data set consisted of 5 herds. Milk production and SCC were recorded during 1 lactation. From approximately 100 does in each herd, milk samples were collected on 3 occasions during lactation for bacteriological culture. Linear mixed regression was used to estimate the effect of IMI on MY. The second data set consisted of 6 large herds, in which some of the goats had an extended lactation (≥2 yr). Milk yield and SCC data were recorded without bacteriological culture. The data showed that bacterial infection was related to an increase in SCC. Infections with major pathogens were rare and associated with a decreased MY; infection with coagulase-negative staphylococci did not affect MY, whereas infection with Corynebacterium bovis was associated with increased MY. A negative correlation was observed between SCC and MY, but the data suggested that this negative correlation was attenuated rather than caused by IMI. Furthermore, SCC seemed to be affected by MY via a dilution effect. Hypotheses about biological mechanisms behind these observations are discussed. This paper shows that MY losses caused by subclinical udder infections are limited in goats, and that SCC cannot be used to estimate the magnitude of these losses.