This is because a lot of the people Kinsey studied were not just mainstream folks, so there was some skewing. If you have too many incarcerated people in your sample, for example, more than in the general population, your results will not reflect the general population.
Not to mention that Kinsey (who was bisexual, and he and his wife had other sexual partners) encouraged his own staff to have wide sexual experiences so they could "feel more comfortable" talking to people about their sex lives.
I don't give his results much credence, frankly.
But while homosexuality isn't rare, it also isn't that common. I still think the number of around 2%, possibly a little less, of mostly exclusively homosexual is about right. I also tend to discount the kind of experimental things we all (or many of us) might have tried when we were young as having a lot of meaning. Hormones can be pushy.