This paper explores the determinants of start-up size by focusing on a cohort of 6247 businesses that started trading in 2004, using a unique dataset on customer records at Barclays Bank. Quantile regressions show that prior business experience is significantly related with start-up size, as are a number of other variables such as age, education and bank account activity. Quantile treatment effects (QTE) estimates show similar results, with the effect of business experience on (log) start-up size being roughly constant across the quantiles. Prior personal business experience leads to an increase in expected start-up size of about 50%. Instrumental variable QTE estimates are even higher, although there are concerns about the validity of the instrument.