The rule of thumb that I am aware of is 1000 ft of annular fill and 10 mins contact time. Need to fulfil both criteria. That should take care of hole size and also pump rate issues. Low pump rate still 1000 ft of fill and 10 mins contact time. High pump rate achievable 10 mins contact time and 1000 ft of fill. This came from a number of experiments on simulated cores in large scale experiments and field evaluation. Of course this is not the whole story with regards to spacer design. For OBM need to have water wetting additives in the spacer system and materials to ensure compatibility between the cement spacer system and mud. As mentioned chemical washes are also effective at breaking up certain mud types. Along with these materials it is also desirable to have density and rheology hierarchy to maximise mud removal. The mud can be tested to determine how hard it can be to remove and the spacer rheology designed to maximise mud removal in the eccentric annulus. So the rheology and density profile of the cement should be greater than the spacer system and the spacer system rheology and density should be greater than the mud system. All of this course within the allowable ECD of the well. So there is much more to spacer systems than 10mins contact time (which is desirable to allow for mud breakup and water wetting the surfaces if OBM is used). The old school idea of pumping in turbulent flow is not always possible within ECD limits and mud systems may not always be easily removed by turbulent flow which may not be achievable anyway. Lab testing and computer simulation in eccentric annuli is the only way to design spacer system to maximise mud removal and maximise the successful isolation required in the well.
I am not a cement guru :) but I know the chemical washer has to allow at least 8 minutes contact time across zone of interest, when in turbulent flow, or a combination of chemical washer and spacer to be used if well control issue do not allow such contact time.