Sample size - statistics problem

[Guest]

A marketing executive wishes to estimate the average monthly sales of a particular design of gold necklace in local shops. He decides that an estimate, correct with 10% of the population average with a probability of 0.9545(2 standard error )shall be considered dependable. How big a sample of shops needed? Assume that, based on a previous enquiry, the coefficient of variation of monthly sales of this design of gold necklace in local shops is 30%.