Also, how many people you need depends heavily on whether the thing you're researching affects a lot of people. If settings has a problem that affects 40% of people, then a sample of 10 people would yield results representative of the whole population.

Sample size is a very weird, often kinda counter-intuitive topic.