Multibrand pricing as a strategy for consumer search obfuscation in online markets

• DOI: http://doi.org/10.1111/jems.12239
• Published date: 01 June 2018
• Date: 01 June 2018

Received: 29 July 2015 Revised: 1 September 2017 Accepted: 1 December 2017

DOI: 10.1111/jems.12239

ORIGINAL ARTICLE

Multibrand pricing as a strategy for consumer search obfuscation

in online markets

Stephen McDonald Colin Wren

NewcastleUniversity Business School,

Newcastle-upon-Tyne, England, UK

(Email: stephen.mcdonald@ncl.ac.uk;

colin.wren@ncl.ac.uk)

Earlierversions of the paper were presented at

the 2013 IIOC and EARIE conferencesand at a

staffseminar at t he Department of Economics,

Universityof Sheffield. The authors are grate-

fulto the par ticipants at these events,including

TimothyRichards, Marco Haan, Joao Montez,

andKonrad Stahl. The authors thank Ilona

Serwickafor assistance, Nils Braakmann,

Jonathan Jones, and John Wildmanfor com-

ments,and t he Nielsen Companyfor supplying

the advertising data, but are responsible forthe

paper'scontent. The aut hors are also grateful

toanonymous referees and a coeditor, whose

commentshave greatly improved the paper.

Abstract

This paper argues that a firm with multiple brands can obfuscate consumer search by

excluding the brands of other firms from a consumer's consideration set. This is exam-

ined empirically by regressing price data for a leading U.K. motor insurance price

comparison site (or “shopbot”). It finds that multibrand firms own three-quarters of

brands in this market, and that allowing for other brand strategies, they post signifi-

cantly lower and clustered prices relative to other firms. The firms also conceal their

brand ownership, consistent with search obfuscation. The results are not otherwise

explained and they have implications for market competitiveness.

1INTRODUCTION

It has long been recognized that consumer search costs give firms market power (Scitovsky, 1950), but it is only recently that

search obfuscation strategies have been considered as a way to endogenize these search costs. To gain a competitive advantage,

the firm can either reduce the effectiveness of consumer search (e.g., Ellison & Wolitzky, 2012; Wilson, 2010) or it can influ-

ence the alternatives that consumers perceive as relevant to a purchase decision (Piccione & Spiegler,2012; Spiegler, 2011). The

interest in search obfuscation reflects the growth of online markets, which has reduced search costs and led firms to strategically

innovate (Ellison & Ellison, 2009). However, despite a burgeoning literature on this topic, there has been little empirical explo-

ration of search obfuscation. To our knowledge,Muir, Seim, and Vitorino (2013) is one of the few studies that have explored this

issue empirically, but this is in the context of an offline market in which there is a nonstandardized reporting of prices and the

consumers incur physical costs in visiting sellers. This paper remedies this deficiency in the literature by investigatingconsumer

search obfuscation empirically for an online market.

Broadly, search obfuscation models are of two types. One stream of the literature assumes that consumers are rational and

can correctly infer all prices, so that it focuses on devices like product “add-ons” that are revealed only at the time of purchase

(Ellison, 2005; Gabaix & Laibson, 2006; Seim, Vitorino, & Muir, 2017). These and other pricing strategies are a strong focus for

research in the marketing literature, which is surveyed by Ahmetoglu, Furnham, and Fagan (2014). The other stream supposes

consumers are boundedly rational, and it considers framing and pricing strategies that either confuse the consumer (Chioveanu &

Zhou, 2012; Seim, Vitorino, & Muir, 2016) or lead to the adoption of simple heuristics (e.g., Spiegler, 2006). In this latter case,

Eliaz and Spiegler (2011) suppose firms use marketing devices to influence the alternatives that the consumers include in their

consideration sets. These sets are used by consumers to reduce the number of options to a manageable size, from which a choice

is made after further research. Consideration sets are recognized as a screening device in the marketing literature (Howard &

Sheth, 1969), and more recently,they are a feature of online markets in which consumers form consideration sets using the lowest

ranked prices or the first page of search results from an Internet comparison site (Masatlioglu, Nakajima, & Erkut, 2012).1The

J Econ Manage Strat. 2018;27:171–187. © 2018 WileyPeriodicals, Inc. 171wileyonlinelibrary.com/journal/jems

172 JOURNAL OF ECONOMICS & MANAGEMENTSTRATEGY

marketing devices considered by Eliaz and Spiegler (2011) include comparative advertising, salespersons, product positioning,

and search engine optimization, but these devices are likely to work much less well in online markets if consumers base their

consideration sets on the lowest-ranked prices.

This paper considers an alternative strategy for influencing the consumer choice in an online market, which is where a firm

uses multiple brands to “load” the consideration set with its own brands and obfuscate consumer search by not revealing its

ownership of these brands. Firm ownership is relevant to the purchase decision because it signals reputation and trust in a brand

(Cabral, 2000), and if it is known, then it might cause a consumer to swiftly discount all but the cheapest brand from the seller

when the product is otherwise similar. Search obfuscation occurs as the concealment of the ownership deceives consumers into

unwittingly excluding the brands of other firms from their consideration sets. It has features of both types of search obfuscation

model, because it involves bounded rationality (consideration sets) and the nondisclosure of relevant information (ownership).

Further, it is feasible as a web site can be created fora new brand at only a modest outlay, whereas the cost of reaching consumers

is also low due to the use of comparison sites. Ellison and Ellison (2009) find that retailers operate multiple sites to sell the same

product range, but their empirical results are about add-ons that are used by firms to lure-in price sensitive consumers to consider

superior products.

This paper explores multibrand pricing as a strategy for search obfuscation by taking data for a price comparison web site

or “shopbot” in which there is a standardized reporting of prices and other policy attributes. The examination is for a leading

U.K. motor insurance price comparison site, which (ironically) is called confused.com. Like other U.K. motor insurance sites,

in response to an online enquiry, the quotes are reported in ascending price order, with information presented on the premium

and other policy attributes. The prices in this market are consumer-specific, and the data are collected for 22 individual types,

by age, occupation, and sex, for eachof five car types, giving 110 “submarkets.” This is at four dates, so that there are 440 “price

sets” and a total of 28,327 observations on prices. These data are examined as a price rank for each price set. The regression

analysis seeks to determine if the multibrand firms post prices that are both low and clustered relative to the single-brand firms.

The paper finds that 16 multibrand firms own three-quarters of the brands in this market. These firms pursue different strategies

with at least some of their brands, including horizontal differentiation and market segmentation, but allowing for these, the

analysis shows that they post prices that are both relatively low and clustered. This is not due to unobserved heterogeneity at

either a brand or firm level, such as from common costs or risks. Rather, as only one firm openly declares its brand ownership, it

is believed to reflect search obfuscation, with firms seeking to increase the probability that their product is chosen. It points to a

new form of strategic behavior that has implications for the competitivenessof online markets. Further, as firms can post pr ices

with ease and firm ownership is opaque in online markets, it suggests that this kind of behavior maybe per vasivein e-commerce.

The structure of the paper is as follows. The next section sets out the hypotheses and considers the identification strategy for

ruling out alternative motives for multibrand pricing. Section 3 describes the U.K. motor insurance market, the confused.com

web site and the data collection. Section 4 considers the unconditional pricing behavior of the multibrand firms, and Section 5

presents the regression results. Finally, Section 6 concludes.

2MULTIBRAND PRICING

To our knowledge, Ireland (2007) is the only paper that formally models multibrand pricing as a strategy for consumer search

obfuscation. Multiple-price posting occurs as pairs of firms pursue joint profit maximization, but do not disclose this to con-

sumers. This is advantageous to the joint-firm as it can capture those consumers who sample both of its prices only, and to whom

it can charge a higher price. The model is stylized, because consumers search randomly at an ex ante fixed sample size of either

one or two prices only, where the prices are the same for the joint-firm. There is no consideration set in this model, but consumer

search is similar because if a firm is not included in the fixed sample, then its price is not transacted. However, there are features

of this model that make it inappropriate for an online market.

In particular, search obfuscation in the Ireland model implies that it is necessary to exclude all other firms from a consumer's

searched sample, but this is infeasible in an online market where a consumer can search a very large number of brands at a low

cost and there is strong competition between firms. Further, in an online market, the consumers do not choose their consideration

sets (or fixed samples) randomly, but rather they use criteria to form these sets (Dulleck, Hackl, Weiss, & Winter-Ebmer, 2011;

Ratchford, 2009). This may involve a ranking of products such as the lowest prices or the first page of results from an Internet

search engine (Masatlioglu et al., 2012). If consumers use the lowest prices on a comparison site to form their consideration

sets, then it is also no longer possible to exploit consumers by charging higher prices. Indeed, the evidence for online marketsis

that sales are distributed among the lowest posted prices (Baye, Rupert, Gatti, Kattuman, & Morgan, 2009), so that a firm must

set its prices low to be included in a consideration set. Under a uniform distribution of prices and a low level of search costs,