Coxiella (C.) burnetii, a zoonotic bacterium, is prevalent in dairy farms. Some cows develop a persistent infection and shed C. burnetii into milk and occasionally by amniotic fluid at calving. Serological diagnosis of Q fever in humans is performed by phase (Ph)-specific antibody tests; PhII antibodies usually indicate an acute infection, while the development of a chronic infection is characterised by elevated PhI antibody titres. Phase-specific tests have now been established for diagnosis of coxiellosis in cattle. Additionally, an interferon-γ (IFN-γ) recall assay has been implemented to assess cellular immunity to C. burnetii in cattle. Milk samples from all lactating cows (n = 2718) of 49 Bavarian dairy farms were collected through a convenience sample and analysed for phase-specific antibodies. Antibody profiles were evaluated by age. Based on the seropositivity of first-lactation cows, three distinct herd profiles were observed: an 'acute' state of herd infection was characterised by a PhI-/PhII+ pattern. The detection of PhI antibodies (PhI+/PhII+) characterised the 'chronic' state, and seronegative results defined the 'silent' state of herd infection. If antibodies had not been detected in multiparous cows, the herd was considered as probably free of coxiellosis. The analysed cattle herds were noted to have an 'acute' (n = 12, 24.5%), 'chronic' (n = 18, 36.8%), or 'silent' state of herd infection (n = 16, 32.6%). Only three farms (6.1%) were classified as 'free' of C. burnetii. The detection of these herd states over a time period of 4 years in one farm indicated that the described states occur in a cyclical manner. Frequently, a wave-like profile was seen, i.e., a circumscribed seronegative age group was flanked by seropositive age groups. In seronegative animals, IFN-γ reactivity was demonstrated. Seroconversion after vaccination was observed by day 7 post-vaccination in chronically infected herds, whereas in the case of silent infection, it started by day 14. These data indicated a pre-existing immunity in seronegative animals in chronically infected herds. Additionally, IFN-γ reactivity was detected in seronegative calves (>3 months) and heifers from chronically infected farms compared to a negative farm. An infection prior to 3 months of age resulted in cellular immunity in the absence of detectable antibodies. An infection around calving would explain this. The aforementioned circumscribed seronegative age groups are, therefore, explained by an infection early in life during active shedding at calving. Based on these results, an endemic cycle of coxiellosis is proposed: Susceptible young heifers get infected by persistently infected cows. Subsequently, shedding of C. burnetii at calving results in infection and then in cellular immunity in offspring. When these calves enter the cow herd two years later, a maximum of herd immunity is achieved, shedding ceases, and new susceptible animals are raised. In an acutely infected dairy farm, the PhI+/PhII+ serological pattern prevailed in second-lactation cows. In this study, stored sera collected since birth were analysed retrospectively. From the earliest seroconversion, the peak of seroconversion took about 33 months. These data suggested a slow spread of infection within herds. The classification of dairy cow herds is a promising basis for further analysis of the clinical impact of coxiellosis.