McConnell's Experimental Brand Choice Data

Abstract

each of the three containers L, M, and P. Despite this, McConnell showed that after an initial of search behavior his customers developed preferences for the different brands. In addition, he examined the strength of these preferences by exposing the customers to promotional activity during the latter half of the experiment (people being offered cash incentives to buy their hitherto least preferred brand). In this note we compare the nature of these brand preferences with those which commonly occur in the market place. This is important since we believe that artificial experiments of this kind will be the main way for learning about consumers' responses to varying marketing inputs (e.g., pricing, product differences, or promotional activities)-in short, of understanding consumer dynamics. Such experimental results must in due course be related to real life and, therefore, we present a comparison of McConnell's brand preferences with normal buyer behavior. A feature of McConnell's experiment was that the subjects were obliged to buy at regular intervals. This allowed McConnell to obtain simple and clear-cut results by using the familiar purchase sequence approach, i.e., analyzing people's successive purchases, call by call. Unfortunately this approach does not facilitate any comparison between the experiment and the real life situation where people buy at widely varying rates. We have, therefore, reanalyzed McConnell's raw data using a time period approach. This has led to the same conclusions as McConnell about search behavior and the development of brand loyalty, but has in addition allowed us to compare the nature of this brand loyalty with the real life situation. In the next section we discuss the two analytic approaches in more detail.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 14 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.