This work presents some strategies to obtain time-dependent concentration profiles and spectra from time resolved spectroscopic data of complex systems. Multivariate curve resolution–alternating least squares (MCR–ALS) has been of particular interest in this field because of the possibility to implement simultaneously soft-modeling constraints and kinetic models on a multibatch resolution. This approach allows circumventing a rank deficiency inherent to time resolved difference data, even in presence of interferent species. The methodology is illustrated using examples from femtosecond photochemistry and biochemistry. Copyright © 2008 John Wiley & Sons, Ltd.