Time-Series Models in Marketing

Abstract

identify three past stages in marketing model building and implementation, review the current status, and provide some intriguing thoughts on how the model-building process may evolve in response to ongoing and anticipated developments in the marketing environment. It is interesting to note that time-series techniques are not mentioned in their review of the past, receive considerable attention in their assessment of the current situation (mainly in the context of the insights these techniques can provide on marketing's long-run effectiveness), and are again absent in their speculations about the future. In this paper, we complement and extend on Leeflang and Wittink's discussion in three ways. First, we discuss why, in the past, time-series techniques received little attention from marketing model builders and users. Second, we show how each of the identified obstacles has recently been attenuated, and review more extensively the recent contributions of time-series applications to marketing science. Finally, we elaborate on some expected future developments in marketing research, and discuss opportunities and challenges to applied timeseries modelers.

Effects of Word-of-Mouth versus Traditional Marketing: Findings from an Internet Social Networking Site

“…If the effects of WOM and marketing variables could vary over time, a logical alternative to the VAR model is a state-space specification with time-varying coefficients (e.g., Dekimpe et al 2006; Kao and Allenby 2005). This gives where the coefficients for WOM and marketing variables become time varying: where the coefficients for WOM and marketing variables become time varying and β1−3t, δ, γ 1–6 , and θ are parameters to be estimated.…”

Synergistic effects of social media and traditional marketing on brand sales: capturing the time-varying effects

“…For example, Kumar et al (2011) find the time-varying effect of market orientation constructs on business performance by allowing the coefficients to be a linear function of discrete times. Moreover, the state space methodologies such as dynamic linear models (DLM) or Kalman filter (e.g., Osinga et al 2010;Sriram et al 2006) make a strong assumption on the underlying states and assume discrete time and discrete state space (Dekimpe et al 2008;Pauwels 2004). Models that pre-specify the shapes of change or that assume the underlying states can have biased results due to misspecification (Bierens and Pott-Buter 1991) and are sensitive to the number of underlying states (Leeflang et al 2009).…”

Truly Maximize the ROI of Sales' and Marketing's Expenditures With Demand‐Driven Planning

“…Since its original formulation, Little's “Sales Response to Advertising Functions” has become increasingly more powerful and more sophisticated. Now referred to as response functions, the key drivers are: Availability of more accurate and complete data on sales (e.g., scanner data at checkout provided by firms like IRI and Nielsen) and tracking of activities (e.g., digital promotions). Vastly improved computing power. Individual promotional elements of total sales and marketing expenditures extended to include more than one element (e.g., print, TV, digital, sales force). The individual sales and marketing elements extended to include econometric ones (e.g., weather/environment, economic, industry trends, and competition). Qualitatively developed response functions now largely replaced by quantitatively developed ones (Dekimpe, Franses, Hanssens, & Naik, ; Hanssens, Parsons, & Schultz, ). …”