For rocks that are actual rocks and not debris or gravel of the same, I might use a quite high value, or optimistic one, for the design. At 75% it is likely you are in such case. For shallow foundations it is almost of no concern, given the usually imparted stresses. But if pile tip, or whatever, in piles, I could go with the higher values because the strength in the rocks is there. This would imply some consideration on the nature of the discontinuities, if complementary measures should be taken or if by whatever the cause you are lead to not be confident on the strength on the rock. You quote one 3 MPa by one evaluation but if you check pile tip strength for a sound gravel at say 15 m depth you may get that strength for pile tip strength maybe even at safety factor 6.
With rock like with wood or cables the safety factors are quite schizoprhenic, all know that the average behaviour is sound, have strength plenty to go to limit strength and yet, because of being critical or, because the risk too much deflection in wood, and, because the wary intelligence of homo sapiens (this that good can't always be true) huge safety factors appear.
The traditional way of designing foundations, allowable compressions, uses a probabilistic approach with moderate safety factors, say 2.5 to 3. Currently, the characteristic strength approach for so unreliable or variable items as rock or even wood would require far higher safety factors, say, 6 to 10. But if you do such thing you may be also seeing -maybe not precisely for rock- huge unlikely settlements that offend the sight of almost anyone that has put something on the skin of mother earth, making almost any kind of structural connection other than loose (no connections) viable. Hence I think it is sound to promote a practice that predicts in accord to the behaviour, and for the rock case, expect that the high strength is there.
