Porovnať metódy

	Scanner Panel Analysis	Brand-Switching Markov Model
Odbor	Marketing	Marketing
Rodina	Regression model	Regression model
Rok vzniku	1983	1992
Tvorca	Peter M. Guadagni & John D. C. Little	Stochastic-choice marketing tradition; codified in Lilien, Kotler & Moorthy
Typ	Disaggregate multinomial-logit brand-choice model	Discrete-time Markov chain of brand purchasing
Pôvodný zdroj	Guadagni, P. M., & Little, J. D. C. (1983). A Logit Model of Brand Choice Calibrated on Scanner Data. Marketing Science, 2(3), 203-238.	Lilien, G. L., Kotler, P., & Moorthy, K. S. (1992). Marketing Models. Prentice Hall. ISBN: 9780135456415
Ďalšie názvy	Scanner Panel Logit, Guadagni-Little Model, Household Panel Choice Model, Loyalty-Variable Logit	Brand Loyalty Markov Chain, Brand-Switching Matrix Model, Stochastic Brand-Choice Model, Markov Brand-Switching Analysis
Príbuzné	3	3
Zhrnutie	Scanner panel analysis models individual households' brand choices using the purchase histories captured by UPC scanner panels, in which the same households are tracked occasion by occasion with the brand chosen and the prices and promotions they faced. The defining method is Guadagni and Little's 1983 multinomial logit of brand choice, the first model to put scanner panel data to serious analytical use. Its signal innovation is the loyalty variable: an exponentially smoothed measure of each household's past brand purchases that enters the utility function and captures persistent brand preference and state dependence. Alongside loyalty, the model includes price, promotion, and brand intercepts, and yields the probability that a household buys each brand on a given occasion. From the fitted model one recovers price and promotion elasticities at the individual level and can simulate how marketing actions shift choices. The framework launched the modern era of disaggregate choice modeling and remains the reference point for scanner-based brand-choice analysis.	The brand-switching Markov model treats a consumer's sequence of brand purchases as a Markov chain, in which the probability of buying a given brand next depends only on the brand bought last. Its central object is the brand-to-brand transition matrix, whose rows record, for buyers of each brand, the probabilities of staying loyal or switching to each competitor on the next purchase occasion. Estimated from panel purchase histories by simple frequency counts, the matrix can be propagated forward to forecast how shares evolve and solved for its steady-state distribution to predict long-run equilibrium market shares. The diagonal of the matrix measures repeat-purchase loyalty while the off-diagonals measure switching, giving managers a structural picture of competitive churn. The model is the classic stochastic-choice representation of brand dynamics and a conceptual precursor to the loyalty variables used in scanner-panel logit models. It is most useful where purchases are frequent, the brand set is stable, and the first-order memory assumption is approximately satisfied.