We would like to know how molecular clouds turn into stellar clusters, and with what efficiency massive stars form in those clusters, since massive stars are the main agents responsible for evolution of the interstellar medium of galaxies, and their subsequent star-formation history. The imprint of ‘precluster’ molecular cloud conditions can be observed, but only in the least evolved, most embedded clusters, necessarily at wavelengths that can penetrate more than 10 visual magnitudes of extinction. Mid-infrared photometric imaging, most recently and extensively from Spitzer, can be used to select young stellar objects in clustered star-formation environments in our Galaxy and nearby galaxies. Relatively sophisticated methods have been developed, but the fundamental principle remains the selection of sources that have excess infrared emission from circumstellar dust. By fitting radiative-transfer models to a source's spectral-energy distribution between ~1 and ~100μm, we constrain the circumstellar dust distribution and evolutionary state. We can explore many things with this protostellar distribution in mass/luminosity and time/evolutionary state. For example we do not see strong evidence for primordial mass segregation in initial studies. We find evidence of primordial hierarchical substructure, greater clustering at the youngest stages, and even imprints of the pre-stellar Jeans scale. We see correlation of the youngest sources with dense molecular clumps and constrain the timescales for chemical processing and dispersal of those clumps. We have only begun to mine the wealth of existing Spitzer, emerging Herschel and soon ALMA data.