Abstract Pollution sources were quantified by the new olf unit in 20 randomly selected offices and assembly halls in Copenhagen. The spaces were visited three times by 54 judges, who assessed the acceptability of the air: (1) while unoccupied and unventilated to quantify pollution sources in the space; (2) while unoccupied and ventilated to quantify pollution sources in the ventilation system; and (3) while occupied and ventilated to determine pollution caused by occupants and smoking. Ventilation rates, carbon dioxide, carbon monoxide, particulates, and total volatile organic compounds were measured, but did not explain the large variations in perceived air quality. For each occupant in the 15 offices there were on average 6–7 olfs from other pollution sources; 1–2 olfs were situated in the materials in the space, 3 olfs in the ventilation system, and 2 olfs were caused by tobacco smoking. The ventilation rate was 25 l/s per occupant, but due to the extensive other pollution sources only 4 l/s per olf. This explains why an average of more than 30% of the subjects found the air quality in the offices unacceptable. The obvious way to improve indoor air quality is to remove pollution sources in the spaces and in the ventilation systems. This will at the same time improve air quality, decrease required ventilation and energy consumption, and diminish the risk of draughts.