The notion of "fair price" is misleading. Without specifics, any answer you get here is going to sound a little like an economics text book.
The questions to ask are:
- How much is this data worth to the potential customers?
- Is anybody else able or prepared to supply equivalent data at a lower cost?
- Could somebody else supply equivalent data in the future if they realized there was a market for it?
(1) gives an upper limit for how much a (rational) customer will pay. The customer will want to make some "profit" on exploiting the data, so the price has to be a lower than this. If the customer can make a risk free 100% profit on exploiting the data, he or she should be very happy. If the risks are higher, or the investment required to exploit the data is considerable, then the amount a rational customer should be prepared to pay is reduced.
(2) assumes that you can't charge more than the market rate - assuming that there is a market. If there isn't one, (1) governs the price you should charge.
Considering (3) may induce you to charge a price lower than (1) so that nobody else thinks this market is worth entering.
A (probably apocryphal) story to illustrate (3). A reporter asked Bill Gates why Microsoft charged so much for Windows. His reply was that a more interesting question was why they charged so little. (The implication being that by charging less than the price which would extract the most revenue, they could keep the market unattractive to new entrants.)
Hope this helps a little: pricing is hard (particularly for a small market) because you have to estimate the answers to (1) and (2) and have limited opportunities to survey or test.
If the three potential customers are competitors, is this data significant enough to auction it?