Statistics New Zealand’s Linked Employer-Employee Data (LEED) contains monthly information on taxable benefit receipt and earnings for all New Zealanders over a five-plus-year period. There is great potential to use this data to understand heterogeneity among the benefit-receiving population, to examine the medium-term outcomes of individuals receiving benefits, and to evaluate the impact of benefit receipt on these outcomes. This paper has three goals. First, it examines different approaches for defining the relevant benefit population to be used for analysing benefit transitions. In particular, it highlights the differences between examining first benefit spells versus randomly selected spells, and between examining individuals with single benefit spells versus those with multiple spells, during the sample period. Second, it summarises employment and benefit receipt patterns for different demographic groups. The variation found in these results illustrates the large heterogeneity found among the benefit-receiving population. Third, it extends the analysis in Hyslop, Stillman and Crichton (2004), which examines the relationship between benefit spell duration and pre-benefit spell employment experiences and post-benefit spell benefit receipt and labour market outcomes. This new analysis controls for a lengthier period of pre-benefit spell employment experiences than the earlier paper, and estimates separate models for each demographic group. These changes allow us to better control for heterogeneity in the benefit-receiving population.