The construction sector is one of the largest single contributors to employment. A wide and varied set of metrics is used by official channels to predict the number of jobs that will be created for a given value of construction expenditure. These values tend to be shrouded in context specificity (ie, a peak workforce, or the number of jobs created for one year) which meet the agenda of the organisation making the announcement. The research reported in this paper reviews the outcomes of five years of research into labour forecasting culminating in the development of the Labour Forecasting Tool (LFT). The research explores three approaches to deriving labour coefficients (the amount of labour per £m of construction value): i) a review of historic data produced by contractors; ii) a theoretical build-up of labour from bills of quantities and iii) an analysis of published UK national statistics. The shortfalls and advantages of each approach are discussed along with the results of a triangulation of the three methods to test the accuracy of the results. Encouragingly, the results point towards a strong agreement between the three approaches. However, whilst the labour coefficient can provide an estimate of the total labour demand in person-years this must be translated to a meaningful measure of the construction jobs created. To do so requires an understanding of the labour flow during a project