Survey Data for performance metrics

Dear Stats Forum,

I have a question about how to use subjective survey data in the development of design and performance targets for a product. I work in an engineering group that would like to incorporate user satisfaction data in the assessment of product targets in the early stages of design. For instance, 50% of the users should report a "5" or better on a 7-point scale assessing their Satisfaction with the product. The number of subjects is quite small (n = 12 - 16). My question is: does this make sense to do? I'm used to summarizing survey data by taking the mean, and not calculating %'s. My concern is that with such small samples, it would be nearly impossible to hit a performance target such as 80% of users must rate "6" or better on a 7-point scale (which is what they really want). Does anyone have any experience with this? Or can you provide a better way to incorporate subjective ratings into target setting.

Thank you.