The 2x2 was what I was originally thinking but the poster said they were interested in the conditional probability of the response given that the user has a need which means you are looking at just one column (or row) of the 2x2 contingency table. An odds ratio only applies if you are comparing the two groups.

Despite having said early I thought you needed an odds ratio, I think Jewel is right.

All you can really say is given someone had a need, they were x% likely to have that need met (or x/y where x is the number who had need met and y is the total number of people who had a need)

(I hope this is what you're saying Jewel, I really don't want to confuse this thread further. If it is not then I'll bow out!)