From question given, the selling of the cake is assumed to be a Poisson process.
Assume further that the mean number of cakes sold per day is lambda.
Let k=5 = number of cakes sold during the day, then the Poisson pmf (probability mass distribution) is given by
P(k)=lambda^k*e^(-lambda)/k!
or
P(5)=lambda^5*e^(-lambda)/5!
=lambda^5*e^(-lambda)/120
If the average number of cakes sold is 4 per day, then
P(5)=4^5*e^(-4)/120
=0.156 is the probability of selling exactly 5 cakes.
The probability of selling 5 cakes or more (i.e. the sixth and subsequent customers will be told to come back the next day) is then
P(k>=5)=1-(P(k=0)+P(k=1)+P(k=2)+P(k=3)+P(k=4)
=1-(0.018316+0.073263+0.146525+0.195367+0.195367)
=0.371163
(for mean number of cakes sold per day = 4 )