How does having samples of value = zero affect the average of a sample size?

a store has an average of 500 customers a day, from a sample of 56 days of data.
How many days can have a measure of zero customers while keeping withing +/-5% of the average?

I'm not sure what data i need to collect or if i can calculate this based on the above information only.

Any tips would help. I am not a statistics major :(.

boundless constraint

I don't understand the question. Do you think you could restate it more clearly?