Proper management of threatened populations requires prior knowledge of population sizes and structures, however, current techniques to gather this information are generally impractical, costly, and can be physically stressful for the animals. Non-invasive methods (e.g. faecal sampling) that can produce high quality and accurate results are better alternatives. Using faecal samples collected from a Svalbard reindeer Rangifer tarandus platyrhynchus population in the winters of 2008 (N = 158) and 2009 (N = 161), we investigated and validated the feasibility of using faecal pellet sizes to differentiate between female calves, yearlings and adults. We found that pellets from adult females were longer than those from calves, and pellets from adults and yearlings were clearly wider than those from calves. With an accuracy of 91% correct classification, we did show that a combination of faecal pellet dimensions (length, width and depth), rather than a single dimension alone, can allow managers to clearly differentiate between age classes if pellets already identified as being from females are used. We also found a positive relationship between live weight and pellet size of the reindeer. Combined with DNA analysis to identify the gender of the animal that produced the faecal pellet, this information may provide important population parameters and be a valuable tool for the monitoring of various ungulate species including wild reindeer.