Q: For a yes or no survey, how many should say "Yes"?

When a question has two possible responses, it is considered dichotomous. Surveys often use dichotomous questions that ask for a Yes/No or Agree/Disagree response. Results from such surveys can help arrive at conclusions relatively quickly and efficiently.

Coming to your question, first, it is important that you have framed your hypothesis before starting the survey. Once you obtain the data, you will need to test your hypothesis to draw a conclusion about your study population (i.e., if the proposed hypothesis accepted or rejected). From your question, it appeared that you plan on hypothesizing after the results are known (how many Yes responses so that the answer is Yes). If a researcher tests a hypothesis but then omits that hypothesis from their report after they find out the results of their test, it is considered post hoc theorizing, which is a questionable practice.

However, if you simply want to know how to analyze the results (irrespective of the number of Yes/No responses), here is what you should do. After you administer the survey and collect all the responses, you will tally the Yes and No responses and divide them by the total number of participants to get the percentages of Yes and No for each question. Next, depending on the sample type, you will need to statistically test your hypothesis (i.e., compare the percentages for Yes and No responses). For such data (categorical variables, non-parametric testing), perhaps Pearson’s chi-square test would be suitable to test whether you should accept or reject your hypothesis.