Jagpal et al. (1979) showed, using multiplicative nonhomogeneous (MNH) sales response functions, that current econometric specifications for measuring marketing mix interactions and/or carryover effects may be structurally restrictive, regardless of their predictive usefulness. They concluded that there is a need for more flexible specifications of the sales response function which provide greater consistency with marketing theory and hence allow more effective control of marketing instruments. This paper proposes the transcendental logarithmic (translog) formulation because it is more general than the MNH function. First, we develop the rationale for the translog specification. In the second part of the paper, we estimate a translog advertisingsales model using the monthly Lydia Pinkham data. The translog results are shown to be consistent with marketing theory, to differ significantly from previous analyses of the same data, and to have useful policy implications.