The forecasting accuracy of market share models using predicted values of competitive marketing behavior

Selling news and advertising space: The economics of Dutch newspapers

“…The number of independent publishers decreased from 34 in 1970 to 21 in 1990. 1970-1975-1982(Alsem et al (1982, 1988). …”

“…Van Ours (1987), however, shows that after a slight change in the model specification the effect of TV advertising on newspaper revenue largely disappears. Alsem et al (1989a) focus on the demand side of the advertising market using a poll among advertisers to predict advertising revenue. Their prediction, a fall in newspaper advertising in 1989, did not prove to be accurate.…”

Building Models for Marketing Decisions

“…Alsem, Leeflang and Reuyl ( 1989) calibrated market share models for nine brands from three markets using bimonthly data. One of their models pertains to brand j in the dried soup market in the Netherlands: We show the estimated parameters of (18.45) and some statistical test results in Table 18.7.…”

“…For example, model users may insist on evidence that a proposed model outperforms some benchmark. , Alsem, Leeflang and Reuyl (1989) and Brodie, Danaher, Kumar and Leeflang (2000) compared the performance of marketing-mix models to that of a naive model, which predicts next period's value for the criterion variable to be this period's actual value. One might argue that little faith should be placed in the parameter estimates, if the proposed model does not outperform a (naive) model that lacks structural characteristics consistent with marketing traditions.…”

Modeling Competitive Reaction Effects

“…There are different opportunities to obtain these forecasts, ranging from naïve methods such as a ct = a c,t -1 or a ct = 1.05a c,t -1 (competitive advertising increases with 5% over time) to methods that use competitive reaction functions. For a more in-depth review see, for example, Alsem and Leeflang (1994), and Alsem et al (1989).…”