Answer:
I suggest method 1 because it is unbiased and systematic.
The second method requires a lot on the initiatives of customers, and likely to have extreme cases only (liked very much or disliked very much).
The third is biased towards teenagers, which may not be the only category of customers who ordered pizzas.
Again, the fourth requires initiative from the customer, so biased towards customers who had something to say.
Explanation: