Glycerol has many roles in crystallization. It mainly increases the solubility of the protein. Other roles include maintaining inter and intra molecular interactions, filling solvent channels, improving crystal packing, cryoprotection etc. However, if your protein is soluble and stable without glycerol then do not add glycerol in your buffers during the initial crystallization screening stage. Once you have identified the crystallization hit condition, then you may try to optimize the glycerol concentration to see its effect on crystals, diffraction and overall resolution.   

I will completely agree with Avinash. If the protein is stable sans glycerol, it does not need glycerol. 

Also, on the contrary Glycerol is believed to make the protein happy and thus flexible. So much flexible that it does not favor crystallization. So, please avoid glycerol for crystallization.  

Santosh. 

Glycerol is a polyalcohol. It maintains proteins in solutions, by increasing their solubility.

 In crystallography, glycerol is the most used cryosolvent because of its antifreeze properties. In liquid water, each molecule is hydrogen bonded to approximately 3.5 other water molecules, while in ice each molecule is hydrogen bonded to 4 others. Glycerol interferes with ice formation by interposing itself within the water hydrogen-bond network. Since lattices in protein crystals are not only stabilized by protein–protein contacts but also by the interactions established by the hydration shells around each protein molecule, altering the hydrogen-bond network that strengthen crystal contacts is likely to affect nucleation and the rate of crystal growth. Thus, glycerol might act as an antinucleation agent in protein crystallization through a mechanism similar to that that prevents ice formation in glycerol–aqueous mixtures.

The effect of glycerol is mostly accounted for by the way it modifies the dielectric constant and refractive index of the solvent. (See: Vera et al., 2011; Sedgwick et al., 2007).

Glycerol is  viscogenic in nature. It tens to increase the viscosity of the solution which in turn will reduce the diffusional motion of molecules. Thus, it prevents precipitation of protein molecules which preceeds aggregation. crystallization is a controlled and slow precipitation of protein solution. Lower concentration of glycerol may effect the process. This has to be experimentally determined and many proteins do crystallize in presence of glycerol. Anyway, all the best .. 

In protein crystallization, small concentrations of glycerol ( 3% typically, but variable) reduce the "showering effect" , this is abundous nucleation preventing the formation of large single crystals. Glycerol and also other small molecular weight PEG and methyl pentane diol, at low concentrations, shield protein surfaces because of their amphiphilic nature, thus preventing too many uncontrolled protein-protein contacts and increasing solubility. Upon addition of a few percentage of glycerol, nucleation and crystal growth are slowed down and crystal packing happens in a much more controlled fashion. This change in kinetics also affect the speed of protein deposition of protein on the respective crystal faces and may favor growth in a crystal dimension that otherwise remains tiny, eg. you may change needle to plate, cubic  or prism shape. This has been demonstrated for lysozyme in the presence of 3% PEG 400. 

I would recommend doing both with and without glycerol if you can make enough protein. If you feel glycerol is interfering, you may lower its percentage (2-5%).  Although you may feel its an additional variable for crystallisation optimization, but if you get crystals with glycerol, you cryoprotectant optimization becomes much simpler. Most likely, then you protein crystals would survive higher glycerol concentrations (typically 30%) used for cryoprotection as that would have already "seen" glycerol.

As mentioned above (by Julie Bouckaert), glycerol can act as a crystallisation retardant. This property was exploited by Cosenza et al. (2000, Acta Cryst D56, 1688-1690) for the crystallisation of Pyruvate Phosphate Dikinase from Trypanosoma brucei. Usually however, glycerol is seen as a poisoning agent for crystallisation for its ability to keep proteins soluble. So it all depends on the particular protein you are working on.

Is Glycerol bad or good?

If you don't need it, don't have it in. If you need it for stability, don't worry (actually, you are free to worry). There are no crystallization data on possible substitutes for glycerol either. 

Glycerol may be useful as a retardant when things grow too fast (problem also often seen in nanodrops?)

file:///H:/crstlization/20160516/cg101364m.pdf

(Practical Use of Glycerol in Protein Crystallization

Laura Vera,† Bertrand Czarny,† Dimitris Georgiadis,‡ Vincent Dive,† and Enrico A. Stura*,†)

https://www.researchgate.net/publication/309041540_Important_Factors_Influencing_Protein_Crystallization

Article Important Factors Influencing Protein Crystallization