Pritchard, John Fabian2011-06-082011-06-082011-06-08http://hdl.handle.net/10539/10044MBA - WBSThis research is concerned with developing a technique for modelling the promotional response of fast moving consumer goods in an environment that is evolving through new product introductions and changing promotional frequency patterns. A technique is developed that achieves this objective and also achieves the objective of differentiating between own brand and category level cross period effects, as well as own brand and category level expansion effects. This is the first public domain technique that the author is aware of that achieves these objectives. Through using a theoretically derived data set and various simulations, the research also demonstrates that significant bias is introduced into the coefficient estimates of extant techniques that do not allow for the introduction of new products and changing promotional patterns when these effects do in fact exist. Through the use of Monte Carlo simulation methods, the reliability of the proposed solution algorithm is investigated in the presence of un-modelled disturbances. For the disturbance domain that was chosen for the research, it is shown that the estimation technique is quite sensitive to the disturbances, particularly when estimating cross period and category expansion effects. Directions for further research are includedenConsumer behaviourThesis