Just as fine precipitate dispersions are sought for high yield strength in alloys, nanoscale, mixed-phase microstructures are sought for high electromechanical performance in oxide ferroelectric materials. Spinodal decomposition is an effective route to fine-scale alloy microstructures and is exploited commercially, for example, in bronzes. Spinodal decomposition has been linked to nanoscale ferroelectric microstructures, but a comprehensive analysis is missing. Here spinodal decomposition in ferroelectric crystals is analysed in 3D including the local electric field using the ferroelectric multi-phase field model for polymorphic phase boundary (PPB) systems. A polarisation consulate temperature Tsp is identified: in single phase ferroelectric, Tsp=Tcw, and, in PPB ferroelectrics, Tcw,ζ=0<Tsp<Tcw,ζ=1. The stability of fluctuations in polarisation depends on temperature, is isotropic in wavevector k→, and corresponds to 2D waves in ceramics. General predictions are summarised and specific predictions for single phase and a PPB system (1−x)Ba(Zr0.2Ti0.8)O3-x(Ba0.7Ca0.3)TiO3 (BZT-xBCT) are made. At 284K, the minimum wavelength is lcritBT=166nm in BaTiO3 and lcritPZT=1.70nm in Pb(Zr,Ti)O3, suggesting that spinodal decomposition is less likely to be observed in BaTiO3. For BZT-40BCT, the minimum observable wavelength at TPPB is ∼22 unit cells lcritBZT−40BCT=8.72nm and presents a viable route to nanoscale rhombohedral + tetragonal microstructures. Domain size engineering by spinodal decomposition with controlled grain structures and quenching in oxide and organic ferroelectrics is identified as a new route to optimise performance.