Telomeres are protective, nucleoprotein structures at the end of chromosomes that have been associated with lifespan across taxa. However, the extent to which these associations can be attributed to absolute length vs. the rate of telomere shortening prior to sampling remains unresolved. In a longitudinal study, we examined the relationship between lifespan, telomere length and the rate of telomere shortening in wild, purple-crowned fairy-wrens (Malurus coronatus coronatus). To this end, we measured telomere length using quantitative polymerase chain reaction in the blood of 59 individuals sampled as nestlings and 4–14 months thereafter, and in 141 known-age individuals sampled on average three times across adulthood. We applied within-subject centring analyses to simultaneously test for associations between lifespan and average telomere length and telomere shortening. We reveal that the rate of telomere shortening and to a lesser extent telomere length in the first year of life independently predicted lifespan, with individuals with faster shortening rates and/or shorter telomeres living less long. In contrast, in adulthood neither telomere shortening nor telomere length predicted lifespan, despite a considerably larger data set. Our results suggest that telomere length measured very early in life (during development) and longitudinal assessments of telomere shortening during the first year of life constitute more useful biomarkers of total life expectancy than either telomere length measured after development, or telomere shortening later in adulthood.