A renewed interest in understanding the geomechanical properties of organic rich mudstones is driven by their importance as unconventional hydrocarbon reservoirs and as sealing units to carbon sequestration sites. Difficulty in extracting good quality core for laboratory determination of mechanical and petrophysical measurements hinders the detail that can be captured in numerical simulations for the purpose of modelling leakage rates. Furthermore, the strong heterogeneity and compositional variation that exist across all scales makes understanding the sensitivities of mechanical properties to changes in composition – particularly the organic fraction – difficult. Numerical homogenisation schemes widely used in composites engineering are in theory well suited to dealing with the compositional and mechanical heterogeneities in mudstones. Novel application of high resolution atomic force microscopy techniques allow for large statistically representative datasets of mechanical properties with nanometre resolution to Jurassic carbonaceous shales of the central USA has provided good quality data on mechanical properties particularly that of organic matter. Moreover, the complexity of the probability density function of young’s modulus for organic matter suggests extreme sensitivity to geochemical variation. Nano indentation techniques have been generally adopted to give small scale measurements of component properties, however rely on the statistical deconvolution of numerous indentation experiments to yield estimates the individual elastic moduli of each component. This technique is reliant on the assumption that the elastic moduli of each component are normally distributed. This technique is inherently reliant on an a priori decision as to the number of phases that will be separated from the total signal. Additionally, the chemical complexity of the organic fraction combined with the results of AFM work on shales suggest that the underlying principles used in deconvolution may not be valid. Comparison AFM to nanoindentation results is paramount to verify consistent measurements over the scale range. Additionally of the results from PeakForce QNM against other AFM techniques has to date only so far be shown to provide consistent outputs on soft biological samples and not verified at high stiffness. Such verification will open the door for the application of a wide range of AFM techniques to mudstones. AFM combined with SEM provides a good way of understanding the complexities of shale microstructure and verifying the applicability of homogenisation schemes – which generally assume a much simpler geometry to the microstructure – using AFM data derived from real mudrocks to populate homogenisation schemes allows comparison against synthetic microstructures generated in multiscale modelling approaches.