To investigate whether coinciding peak serum caffeine concentration with the onset of exercise enhances subsequent endurance performance.Randomised, double-blind, crossover.In this randomised, placebo-controlled, double-blind crossover study, 14 male trained cyclists and triathletes (age 31±5year, body mass 75.4±5.7kg, VO 2 max 69.5±6.1mLkg −1 min −1 and peak power output 417±35W, mean±SD) consumed 6mgkg −1 caffeine or a placebo either 1h (C 1h ) prior to completing a 40km time trial or when the start of exercise coincided with individual peak serum caffeine concentrations (C peak ). C peak was determined from a separate ‘caffeine profiling’ session that involved monitoring caffeine concentrations in the blood every 30min over a 4h period.Following caffeine ingestion, peak serum caffeine occurred 120min in 12 participants and 150min in 2 participants. Time to complete the 40km time trial was significantly faster (2.0%; p=0.002) in C 1h compared to placebo. No statistically significant improvement in performance was noted in the C peak trial versus placebo (1.1%; p=0.240). Whilst no differences in metabolic markers were found between C peak and placebo conditions, plasma concentrations of glucose (p=0.005), norepinephrine and epinephrine (p≤0.002) were higher in the C 1h trial 6min post-exercise versus placebo.In contrast to coinciding peak serum caffeine concentration with exercise onset, caffeine consumed 60min prior to exercise resulted in significant improvements in 40km time trial performance. The ergogenic effect of caffeine was not found to be related to peak caffeine concentration in the blood at the onset of endurance exercise.