It's difficult to tie changes in customer retention to not having a doorman in a "hard data" way. Ideally you'd want to do an A/B test of doorman vs non-doorman, but you'd need multiple hotels for that to work.