Up through the 19th century, many North Americans participated in the tradition of Christmas "side hunts", in which they competed at how many birds they could kill, regardless of whether they had any use for the carcasses and of whether the birds were beneficial, beautiful, or rare. At the end of that century the U.S. ornithologist Frank Chapman, an officer in the recently formed National Audubon Society, proposed counting birds on Christmas instead of killing them.
In 1900, 27 observers took part in the first count in 25 places in the United States and Canada, 15 of them in the northeastern U.S. from Massachusetts to Philadelphia. Since then the counts have been held every winter, usually with increasing numbers of observers. For instance, the 101st count, in the winter of 2000-2001, involved 52,471 people in 1,823 places in 17 countries (but mostly in the U.S. and Canada). The Audubon Society now partners with Bird Studies Canada, the Gulf Coast Bird Observatory of Texas (responsible for CBCs in Mexico), and the Red Nacional de Observadores de Aves (RNOA, National Network of Bird Observers) and the Instituto Alexander von Humboldt of Colombia.
Each individual count is performed in a "count circle" with a diameter of 15 miles or 24 kilometres. At least ten volunteers, including a compiler to manage things, count in each circle. They break up into small parties and follow assigned routes, which change little from year to year, counting every bird they see. In most count circles, some people also watch feeders instead of following routes.
Counts can be held on any day from December 14 to January 5 inclusive.
The results are by no means as accurate as a human census. Not all the area even in the count circles is covered, and not every bird along the routes is seen or identified. Big flocks can't be counted precisely. Also, telling whether a bird has been counted twice can be difficult. The rules address this problem by prohibiting counting birds when retracing one's route, except for species that the party hasn't seen before. Also, when a large roost of some species occurs in a count circle, an expert estimates the number for that species during the morning or evening and usually no individuals are counted at other times. Observers can attempt to keep track of flocks of mobile birds such as crows, and can use their judgement, even sometimes recognizing an individual bird or at least recognizing that two birds of the same species are different individuals.
The results, providing data on winter ranges of birds, are complementary to those of the Breeding Bird Surveys.
Participation is open to all. Observers pay a $5 fee (except feeder watchers, U.S. participants under 19 years old, and Latin Americans in their home countries). The fee supports compilation and publication of the data. U.S. participants who pay or who are 18 or under receive a copy of the issue of American Birds that summarizes the results and includes articles on trends and regions.