What percentage do you take? What levels do you think the market will reach for, say, qualified mentors in <X>?

When you get real numbers, there's a lot you could do with this data. For example, if Ruby tutors make 20% more than Python tutors, there's an economic signal there. While the raw data might seem limited in use, there are a lot of open questions about the programming economy that one could start answering.

Especially powerful would be to capture code from remote pairing sessions and get some quality-labelled samples (labels by tutor price level) and perhaps (many years later) attack the automated code-quality assessment problem (on which well-labelled data is very rare).