Incorporating Consideration Sets Into Models of Brand Choice

Citation:

Prakash Nedungadi and Vinay Kanetkar (1992) ,"Incorporating Consideration Sets Into Models of Brand Choice", in NA - Advances in Consumer Research Volume 19, eds. John F. Sherry, Jr. and Brian Sternthal, Provo, UT : Association for Consumer Research, Pages: 251-252.

Advances in Consumer Research Volume 19, 1992      Pages 251-252

INCORPORATING CONSIDERATION SETS INTO MODELS OF BRAND CHOICE

Prakash Nedungadi, University of Toronto

Vinay Kanetkar, University of Toronto

There has been a recent resurgence of interest in the notion of the consideration set and its precise role within the brand choice process. Behavioral researchers have examined the process of consideration set formation and use (Alba, Hutchinson & Lynch 1991; Baker, Hutchinson, Moore & Nedungadi 1986; Nedungadi 1990). A few quantitative models of consideration set formation have also been developed (e.g., Hauser & Wernerfelt 1990; Roberts & Lattin 1990). In this paper, we propose a model of brand consideration and choice, and examine the implications of incorporating consideration into brand choice models.

CONSIDERATION SETS

In order to incorporate consideration sets into the choice process, we view choice in two stages -- brand consideration and brand selection (see Nedungadi 1990). In the first stage, internally and externally generated retrieval cues lead to a situation-specific consideration set being brought to mind. In the second stage, the brands in this set are compared and one brand is selected for purchase. This separate focus on the consideration and selection stages is useful for a number of reasons. First, it highlights the fact that the processes that lead to brand consideration may be quite different from those that result in choice from the consideration set. While researchers have often alluded to two- stage choice processes, they have essentially treated consideration as an extended version of choice. Thus, brand preferences are often believed to drive both consideration set formation and choice. However, Nedungadi (1990) presents evidence that factors other than brand preference, such as brand accessibility and external retrieval cues may drive consideration set formation during choice.

A separate focus on brand consideration also allows for a more dynamic conceptualization of consideration sets. Thus, consumers need not "have" consideration sets that they carry to each purchase occasion, instead, the consideration set could be "brought to mind" spontaneously, as a first step in choice. Finally, a separation between consideration and selection stages also facilitiates a distinction between the effects of marketing variables at each stage.

MODEL

We propose a simple logit model to capture the first, consideration stage of the choice process. Although logit models have been used extensively in the choice context, they have not been applied to brand consideration. Let c_h ΠU be the set of brands considered by individual h, and U be the universe of all brands in the product class. We assume that c_h contains at least one brand for all individuals. In addition, consistent with Hauser & Wernerfelt (1990) and Roberts & Lattin (1990), we assume that the probability of considering each brand is independent. Let A_ih = 1 if brand i is contained in the consideration set of individual h, c_h, 0 otherwise. Then, the probability of considering brand i is expressed as:

EQUATION

where w_ih is the consideration "value" of brand i for individual h. If there are M explanatory variables, then consideration value may be written as a linear function of these explanatory variables. Functionally, we expect that w_ih = -_m=1DBA12()^M g_miZ_mih where Z_mih is the value of the m^th explanatory variable and g_mi is the parameter associated with that variable for alternative i and is to be estimated empirically. Further, the value of g_mi is indicative of the effect of the explanatory variable on likelihood of considering the brand.

Further, let y_ih = 1 if brand i is chosen by individual h, 0 otherwise. Then the probability of choosing brand i from consideration set c_h could be expressed as:

EQUATION

where v_ih and v_jh are brand utilities for brands i and j respectively, for individual h. This simple two-stage choice model is different from the traditional multinomial logit model where probability of choice is directly expressed as:

EQUATION

In this traditional model, brand choice probabilities are determined by relative utilities alone.

Thus, we propose a simple, two-stage choice model that incorporates both subprocesses of brand consideration and brand selection. The probability of brand consideration is expressed as a function of consideration value, while the probability of selection is expressed as a function of relative preference for brands in the consideration set. We test this model on a set of experimental data that includes measures of brand accessibility, consideration, choice and preference. Nedungadi (1987) collected data on choices of fast-food outlets. PC/LIMDEP (Greene 1988) was used to estimate the consideration choice model on this experimental data and compare it to the traditional choice model, both in terms of prediction and diagnostic insights. While details of the test are provided in Nedungadi & Kanetkar (1991), our test led to the following conclusions:

1. Brand accessibility and acceptability have a greater influence on the brand consideration stage of choice and a lesser influence on the brand selection stage. As a result, marketing influences (e.g., reminder advertising) that increase accessibility are most important at this stage of the choice process.

2. While brand preferences are important during brand consideration, their chief role is in the brand selection stage when brands are compared for choice. Again, marketing influences that affect brand utilities (e.g., price discounts) are most likely to act at this stage, once brands have entered the consideration set.

3. Incorporation of the brand consideration stage greatly improves the predictive power of brand choice models.

REFERENCES

Alba, Joseph W., J. Wesley Hutchinson, and John G. Lynch, Jr. (1991), "Memory and Decision Making," in Handbook of Consumer Behavior, eds. Thomas Robertson and Harold Kassarjian, New York: Prentice Hall, 1-49.

Baker, William, J. Wesley Hutchinson, Danny Moore, and Prakash Nedungadi (1986), " Brand Familiarity and Advertising: Effects on the Evoked Set and Brand Preference," in Advances in Consumer Research, Vol. 13, ed. Richard J. Lutz, Provo, UT: Association for Consumer Research, 63

Greene William (1988), LIMDEP, Econometric Software Inc., New York.

Hauser, John R. and Birger Wernerfelt (1990), "An Evaluation Cost Model of Consideration Sets," Journal of Consumer Research, 16(March) 393-408.

Nedungadi, Prakash (1990), "Recall and Consumer Consideration Sets: Influencing Choice Without Altering Brand Preference," Journal of Consumer Research, 17(December), 263-276.

Nedungadi, Prakash and Vinay Kanetkar (1991) "Modeling Brand Consideration and Choice: Tests on Experimental and Scanner Data," Working Paper, University of Toronto, Toronto, Canada.

Roberts, John H. & James M. Lattin (1990), "Development and Testing of a Model of Consideration Set Formation," Working Paper 90-014, Kensington, NSW, Australia: Australian Graduate School of Management, (April).

----------------------------------------