A Review and Conceptual Framework for Understanding Personalized Matching Effects in Persuasion - Pablo Briñol
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
A Review and Conceptual Framework for Understanding Personalized
Matching Effects in Persuasion
Jacob D. Teeny Joseph J. Siev
Northwestern University The Ohio State University
Pablo Bri~
nol Richard E. Petty
Universidad Autonoma de Madrid The Ohio State University
Invited and Accepted by Angela Lee, Associate Editor
One of the most reliable and impactful methods for enhancing a persuasive appeal is to match an aspect of
the proposal (i.e., its content, source, or the setting in which it is delivered) to an aspect of the consumer
receiving it. This personalized matching in persuasion (also called tailoring, targeting, customizing, or personal-
izing) comprises a robust and growing literature. In the present review, we describe different types of persua-
sive matches, the primary characteristics of people who are targeted, and the key psychological mechanisms
underlying the impact of matching. Importantly, although most research on personalized matching has con-
cluded that matching is good for persuasion, we also describe and explain instances where it has produced
negative (i.e., “backfire”) effects. That is, more than just the conclusion “matching is good” that many
researchers have drawn, we analyze when and why it is good and when and why it can be ineffective—insight
that can benefit marketers and consumers alike in understanding how personally matched appeals can impact
attitudes and ultimately behavior.
Keywords Attitudes and persuasion; affect and emotion; goals and motivation; personality;
communication
Introduction private information, creating matched appeals has
become more actionable than ever.
Today, modern technology allows one of the most
Although matching has a long history in the
effective methods of persuasion to be implemented
marketing, health, political, communications, and
relatively easily—personalizing messages to the
social psychological literatures, social media and
audience. This technique was first recorded by Aris-
the online revolution have made it an increasingly
totle (Rhetoric, 1.11), and centuries later, a prodi-
important topic of contemporary research and dis-
gious number of studies have shown that whether
cussion (Dijkstra, 2008). With an abundance of data
the personalization comes through the content of
on every active internet user, the potential to create
the persuasive appeal, the source of the appeal, or
messages matched to consumers’ personal charac-
the setting in which the appeal is delivered,
teristics has produced some remarkable effects. For
“matching” some aspect of the communication to
example, several analysts report that the use of U.S.
some aspect of the recipient is one of the most reli-
voters’ personal data to match online appeals to
able and impactful methods of enhancing persua-
their personality traits helped influence the result of
sion (Carpenter, 2012; Noar et al., 2007; Petty et al.,
the 2016 U.S. presidential election (Rathi, 2019).
2000; Rothman et al., 2020). As companies gain
Indeed, with algorithms becoming increasingly
increased access to consumers’ public as well as
accurate in assessing consumers’ characteristics, and
the ease with which artificial “bots” can then use
Received 27 November 2019; accepted 1 September 2020
Available online 15 October 2020 this information to send out matched messages,
The authors thank members of the Group for Attitudes and consumers’ personal data are now regarded by
Persuasion at Ohio State for feedback on earlier versions of this some as “the world’s most valuable resource”
paper. The authors have no conflicts of interest to report regard-
ing this review.
Correspondence concerning this article should be addressed to
Richard E. Petty, Distinguished University Professor of Psychol- © 2020 Society for Consumer Psychology
ogy, 1835 Neil Avenue, The Ohio State University, Columbus, All rights reserved. 1057-7408/2020/1532-7663
OH 43210, USA. Electronic mail may be sent to petty.1@osu.edu. DOI: 10.1002/jcpy.11982 Teeny, Siev, Bri~
nol, and Petty
(Parkins, 2017). Thus, the present time seems ideal illustrative review of studies taken largely from the
for a review aimed at illuminating when and espe- psychological, consumer, and marketing literatures.
cially why matched appeals can be so effective. Because the relevant body of work on matching is
To begin, it is first useful to identify the four vast, however, we cannot cover everything of
classic factors of a persuasion context: the recipient potential interest in these disciplines, let alone
of the communication, the message itself, the source beyond them. Nonetheless, the present work does
of the message, and the setting in which the mes- not restrict itself to a specific consumer characteris-
sage is delivered (Lasswell, 1948; McGuire, 1969). tic (e.g., matching the functional basis of an atti-
Matching is a procedure whereby one of these fac- tude; Carpenter, 2012) nor to a specific content
tors aligns with another. Perhaps the most common domain (e.g., matching in health domains; Rothman
form of matching—and the focus of this review—is et al., 2020; Lustria et al., 2013). Moreover, it is the
an alignment between some aspect of the message first to propose a general framework for under-
recipient and one of the other factors, which we standing the wide variety of consumer characteris-
refer to as personalized matching. This type of tics shown to elicit personalized matching effects
“matching to people” has also been called segment- and the core psychological processes underlying
ing, customizing, targeting, and tailoring (e.g., Haw- them.
kins et al., 2008; Webb et al., 2013).
It is important to emphasize that personalized
matching is only one form that can occur. For
Personalized Matching
example, instead of matching the source to the
recipient (e.g., a female source delivering a message In organizing our review, we categorize the rele-
to a female audience), the source might be matched vant work by the persuasion factor that was
in some way to the message itself (e.g., a female matched to a characteristic of the recipient: (a) the
source delivering a message using female-relevant message content, (b) the source of the message, and
metaphors). In this latter form of matching, because (c) the setting (context) in which the message was
no information about the recipient is involved, we delivered. Because the literature on personalized
label it nonpersonalized matching. In our review, we matching generally shows that it enhances persua-
focus on personalized matching, not only because sion, the following sections outline the diverse vari-
this is the most common form examined in the liter- ables that have produced these positive persuasion
ature, but also because this type of matching serves effects. Later, we outline when and why matching
as the prototypical one employed by marketers. can backfire.
Figure 1 provides a summary and structure of
the ensuing review. We first unpack the various
Message-to-Recipient Matches
ways in which personalized matches can be pro-
duced, describing various types of personalized The most common type of personalized match-
matches that have shown effects. We then describe ing examined in the literature occurs between the
the positive and negative meanings these matches message content and the recipient where some
can assume. We subsequently organize the various aspect of the message is made to align with a tem-
mechanisms by which matching can influence atti- porary or chronic aspect of the recipient. Some
tudes within a classic framework for persuasion research, however, has shown it is also possible to
effects—the elaboration likelihood model (ELM; modify the recipient’s momentary state to match
Petty & Cacioppo, 1986). Finally, we describe how the message (e.g., via priming; Bayes et al., 2020; Li,
marketers can come to expect whether a match will 2016; Loersch et al., 2013; Wheeler et al., 2008).
produce short-term or long-term persuasion conse- Creating message-to-person matches can be as sim-
quences as a function of the process through which ple as using the individual’s name in the address of
the match operates. Although these sections com- the appeal (Howard & Kerin, 2011; Sahni et al.,
bine to provide a comprehensive examination of 2018). These minimal matches have been referred to
documented matching effects and their relevant as placebo tailoring because customization is implied
processes, throughout this review we also present without changing the substance of the appeal
potential avenues for future research, highlighting (Webb et al., 2005). Matching can also be more
not only when matching can be effective, but also encompassing such as when the message includes
when it can be counterproductive. multiple personal characteristics including the per-
Although the present review is intended to be son’s name along with age, brand of product used,
comprehensive, it is not exhaustive. We offer an and so forth (Dijkstra, 2005), or describes multiplePERSONALIZED MATCHING IN PERSUASION 3
Figure 1. Depiction of persuasion processes for matching effects at different levels of message elaboration.
relevant behaviors in which the person has previ- experience higher levels of arousal, they respond
ously engaged (Kreuter & Wray, 2003). more favorably to messages containing more arous-
Research has identified a wide array of recipient ing content (e.g., louder advertisements; Yan et al.,
variables to which a message can be matched. 2016; or ads touting exciting rather than relaxing
Below, we organize these characteristics into the attributes; Rucker & Petty, 2004). Moreover, mes-
following categories: individuals’ affective and cogni- sages promoting products that are themselves
tive states, their goals and motivational orientations, matched to arousal levels (e.g., energy drinks for
their attitude bases and functions, their identities and high-arousal consumers) produce similar outcomes
personality, and their cultural orientation. The cate- (Di Muro & Murray, 2012). Two types of arousal
gories are arranged from relatively low- to more have been identified (i.e., energetic vs. tense; Teeny
high-level characteristics, which we further subdi- et al., 2020), and persuasive attempts for products
vide based on a structure largely devised for orga- matched to the motivational drive associated
nizational convenience. with each can enhance attitude change (Fan et al.,
2015).
Emotive Valence and Specific Emotions. Research
Matching to Affective and Cognitive States
depicts individuals’ more complex affective states
People regularly experience both transient and through a combination of their existing arousal and
longer-term affective states (i.e., feelings/emotions) emotive valence (cf., Russell, 2009), and matching the
and cognitive states (e.g., thinking styles/mind- message’s frame to that valence (i.e., the broad pos-
sets). Depending on the context, different affective itivity or negativity of the person’s affective state)
and cognitive states will be more or less salient (cf., can enhance persuasion. For example, focusing a
Grigorenko & Sternberg, 1995; Keltner & Lerner, message on the positive aspects of making the pur-
2010), making them more or less viable targets for chase (vs. the negatives of not making the pur-
matched appeals. chase) work better for people in a positive (vs.
Psychological Arousal. One aspect across indi- negative) affective state—and vice versa (Cho &
viduals’ affective states is their degree of psychologi- Choi, 2010; Wegener et al., 1994). Because valence
cal arousal (i.e., the subjective sense of energy itself is a binary dimension, though, targeting the
mobilization; Teeny et al., 2020). When consumers specific emotion elicited by the combination of the4 Teeny, Siev, Bri~
nol, and Petty
person’s valence and arousal can produce even
Matching to Goals and Motivational Orientations
more precisely targeted appeals. Whether the per-
son is feeling romantic or fearful (Griskevicius Consumers vary in their salient goals (i.e.,
et al., 2009), sad or angry (DeSteno et al., 2004), or desired endpoints; Fishbach & Ferguson, 2013) as
cheerful or tranquil (Bosmans & Baumgartner, well as their broader motivational orientations (i.e.,
2005), personalizing messages to specific emotional the strategy to pursue goals; Bargh et al., 2010). Dif-
states can enhance persuasion. For example, if a ferent situational and dispositional factors will
person is feeling sad, pointing to the sad (rather make certain goals or motivational orientations
than angering) consequences that a product could more or less salient (cf., Bargh et al., 2010; Moskow-
allay could improve the appeal’s effectiveness. itz & Gesundheit, 2009), which, once known, can be
Active Thinking Style or Mind-set. As with targeted for enhancing one’s persuasive appeal.
affect, matching the content of the message to con- Visceral Drives. The most basic goals con-
sumers’ salient cognitive states can enhance persua- sumers are driven to satisfy are their visceral drives
sion. For example, using more abnormal (vs. (e.g., hunger and fatigue; Loewenstein, 1996). These
normal) appeals for those in a creative mind-set psychobiological needs, though transitory, can be
(Yang et al., 2011) and using arguments that targeted. When an appeal is matched to a con-
emphasize competence (vs. warmth) for those in a sumer’s salient drive (e.g., the energizing benefits of
high (vs. low) power mind-set (Dubois et al., 2016) a product are emphasized to tired consumers), it
have been shown to increase effectiveness. Addi- can enhance the appeal’s effectiveness (Karremans
tionally, using metaphors that match consumers’ et al., 2006; Risen & Critcher, 2011).
cognitive representation of the advocated topic Hedonic and Utilitarian Purchasing Goals. In a
(e.g., advertising antidepressants with metaphors marketing context, one particularly important vari-
that describe depression as “feeling down” for able is consumers’ purchasing goals: whether they
those who cognitively represent depression as are shopping for hedonic (i.e., pleasure-based) ver-
“down” vs. “up”) can enhance persuasion (Keefer sus utilitarian (i.e., functionality-based) products
et al., 2014; Landau et al., 2018). (cf., Abelson & Prentice, 1989). Those with hedonic
Psychological Construal. Construal level refers goals respond more favorably to advertisements
to the extent to which a consumer is in an abstract emphasizing hedonic benefits (e.g., a candle’s aro-
mind-set (i.e., focused on objects’ superordinate and matic and relaxing effects), whereas those with util-
central features) versus a concrete mind-set (i.e., itarian goals respond more favorably to utilitarian
focused on objects’ subordinate and specific fea- benefits (e.g., a candle’s cleansing or bug repellent
tures; see Trope and Liberman (2010) for a review). effects; Chitturi et al., 2008; Klein & Melnyk, 2014).
Accordingly, receiving a message focused on the Other work has also shown how specific types of
abstract desirability (vs. the concrete feasibility) of a message content, such as assertive (vs. nonassertive)
product can produce more positive outcomes for claims for hedonic (vs. utilitarian) shoppers, can uti-
those in an abstract (vs. concrete) mind-set (e.g., lize personalized matching (i.e., hedonic shopping
Fujita et al., 2008; Han et al., 2016). Similarly, when involves impulsive purchasing which matches
a person in an abstract mind-set receives a message assertive statements, like “Just do it”; Kronrod
framed in terms of distant benefits, or a person in a et al., 2011).
concrete mind-set receives one framed in terms of Approach–Avoidance Motivation. In addition to
more proximal benefits, persuasion is enhanced currently activated goals, consumers’ broader moti-
(Jeong & Jang, 2015; see also Wan & Rucker, 2013). vational orientation is also a characteristic which
Related research extends the construal-matching can be matched. For example, people can be more
phenomenon to the resource being requested: situationally or chronically approach-oriented (i.e.,
requests of time (a more abstract resource) for peo- more responsive to incentives) or avoidance-oriented
ple in abstract mind-sets and requests of money (a (i.e., more responsive to disincentives; Carver &
more concrete resource) for people in concrete White, 1994) and matching messages to these orien-
mind-sets (MacDonnell & White, 2015). Other tations can enhance persuasion (Gerend & Shep-
research has extended it to the type of product herd, 2007; Mann et al., 2004). For example, Jeong
being marketed, too (i.e., eco-friendly products are et al. (2011) found that donations to a university
more appealing to consumers in abstract vs. con- increased when approach-oriented people received
crete mind-sets, because protecting the environment a matched appeal in terms of rewards (e.g., your
is a relatively abstract, future-focused initiative; donation helps to expand the number of library
Reczek et al., 2018). books) and when avoidance-oriented peoplePERSONALIZED MATCHING IN PERSUASION 5
received one in terms of punishments (e.g., without consumers are more persuaded by concrete mes-
your donation, the library would have to reduce its sages (i.e., abstract messages inform promotion-fo-
books) versus receiving a mismatched appeal. cused consumers about multiple options for
Regulatory Focus. Probably, the most exten- attaining their goal, whereas concrete messages
sively researched variable in this category is how inform prevention-focused consumers about the fea-
messages can be matched to a person’s regulatory sibility of attaining their goal; Lee et al., 2009;
focus (i.e., how people approach good and avoid Semin et al., 2005; see also Malaviya & Brendl,
bad outcomes, Higgins et al., 2003). According to 2014).
this work, consumers can be more promotion-focused
(attuned to approaching gains and avoiding non-
Matching to Attitude Bases and Functions
gains) or prevention-focused (attuned to avoiding
losses and approaching nonlosses). Like their Naturally, the more a person’s attitudinal posi-
approach/avoidance orientation, consumers’ regu- tion (i.e., their positive or negative evaluation of an
latory focus can manifest dispositionally (Cesario object) matches the attitude expressed by a mes-
et al., 2013), or in response to situational factors. In sage, the more favorably they will respond to it
either case, research clearly shows that matching (Petty & Cacioppo, 1990; see Clark et al., 2013). This
the content of a message to regulatory focus tends does not mean that attitude change is always
to produce positive persuasion outcomes (i.e., greater for pro (vs. counter)-attitudinal appeals,
referred to as regulatory fit; Cesario et al., 2013; Kim, because if the message advocacy perfectly matches
2006; Fransen et al., 2010; for reviews, see Motyka the recipient’s existing attitude, for example, there
et al. (2013), Rothman et al. (2020)). is little room to change toward the message.
In one illustrative study, Lee and Aaker (2004) Beyond the valence of a person’s attitude, then, the
had participants read an advertisement for Welch’s attitude’s underlying basis (Rosenberg & Hovland,
grape juice that either emphasized the benefits of 1960) as well as the attitude’s function (Katz, 1960)
consuming it or the costs of not consuming it. Par- can serve as critical variables in personalized
ticipants induced to have a promotion focus had matching, as we describe next.
more positive brand attitudes following the ad that Affective–Cognitive Bases. The most studied
emphasized the benefits of purchasing, whereas attitude basis involves the affect and cognition con-
participants induced to have a prevention focus tributing to the person’s attitude (Clarkson et al.,
were more positive when the ad emphasized the 2011; Edwards & von Hippel, 1995; Keer et al.,
costs of not purchasing. In addition to increasing 2013; See et al., 2008; for reviews, see Maio et al.,
positive evaluations, matching a message to con- 2019; Petty et al., 2019). Affective bases refer to the
sumers’ regulatory focus can also dampen negativ- feelings and emotions (e.g., anxiety) that underlie
ity in response to unfavorable messages. For one’s attitude, whereas cognitive bases refer to the
example, in denying an employee’s request, when reasons and attributes (e.g., usefulness) that under-
employers used language that matched (vs. mis- lie one’s attitude (Crites et al., 1994). Attitudes vary
matched) the employee’s regulatory focus, the refu- in the degree to which they are based primarily on
sal was better received (Fransen & Hoven, 2013). affect or cognition with attitudes in some domains
Other work has documented more indirect vari- tending to elicit one basis over the other (e.g., affect
ables that can similarly match to regulatory focus. for experiential domains; Breckler & Wiggins, 1989).
For example, promotion-focused consumers Regardless of the specific topic, affective mes-
respond more favorably to advertisements empha- sages (e.g., appeals that feature emotion-evoking
sizing the supply aspect of scarcity, whereas pre- anecdotes) tend to be more persuasive for attitude
vention-focused consumers respond more favorably objects that have an underlying affective basis,
to those emphasizing the demand aspect of scarcity whereas cognitive messages (e.g., appeals that fea-
(Ku et al., 2012). In other research, promotion-fo- ture factual information) tend to be more persua-
cused consumers respond more favorably to higher sive for objects having a cognitive basis (e.g.,
risk, higher reward appeals (consistent with their Edwards, 1990; Fabrigar & Petty, 1999). This
focus on approaching gains and avoiding non- enhanced persuasive effect appears robust, docu-
gains), whereas prevention-focused consumers mented across various domains, such as food con-
respond more favorably to lower risk, lower reward sumption (Dube & Cantin, 2000), movie reviews
appeals (Updegraff & Rothman, 2013). Furthermore, (Mayer & Tormala, 2010), medical self-checks (Mil-
promotion-focused consumers are more persuaded lar & Millar, 1990), and binge drinking (Keer et al.,
by abstract messages, whereas prevention-focused 2013).6 Teeny, Siev, Bri~
nol, and Petty
Although prior research has broadly supported environment because they oppose harming nature,
the notion that affective–cognitive matching whereas others might support it because they want
increases a message’s impact, a very small number to maintain the purity of nature. Matching a mes-
of studies have claimed evidence for mismatching in sage to the relevant moral foundation tends to
this domain (e.g., Millar & Millar, 1990). In these increase relevant attitudinal and behavioral change
instances, it appears the messages have been both (Feinberg & Willer, 2015; Wolsko et al., 2016).
counterattitudinal and, critically, constructed of Similar to affective/cognitive attitude bases,
generally noncompelling arguments. Here, the moral foundations can vary as a function of the
matched messages produced a greater degree of specific attitude topic, or people can more generally
counterarguing against the weak arguments, which vary in their endorsement of a subset of moral
resulted in the matched message’s reduced effec- foundations (Graham et al., 2009). For example,
tiveness relative to the mismatched message. (Later, politically liberal individuals tend to put a greater
we provide detail on the role of strong versus weak emphasis on the care and fairness dimensions,
arguments in matching effects in the section titled whereas politically conservative individuals tend to
Matching Increases Message Elaboration.) emphasize the loyalty, authority, and sanctity
In addition to matching messages to specific atti- dimensions (Haidt, 2012). Thus, when these specific
tudes varying in their basis, some research has moral foundations are emphasized in messages to
focused on the fact that people can differ disposition- those who dispositionally endorse the matched
ally in their attitude bases. Some people tend to have moral basis, it can increase the message’s persua-
more affectively-based attitudes toward a wide vari- siveness (Day et al., 2014; Kidwell et al., 2013;
ety of objects (those high in need for affect; Maio & Voelkel & Feinberg, 2018).
Esses, 2001), whereas others tend to have more cog- Political Bases. The bases of individuals’ atti-
nitively-based attitudes (i.e., those high in lay rational- tudes can also differ as a function of their political
ism, Hsee et al., 2015; or need for cognition, Cacioppo liberalism/conservatism, where each ideology tends
& Petty, 1982). Matching messages to these individ- to correspond to different underlying values (Jost,
ual differences has also been shown to enhance per- 2017; Janoff-Bulman, 2009). Thus, when people with
suasion (Haddock & Maio, 2019). For example, when conservative or liberal beliefs receive appeals that
consumers were advertised a new drink, those who match the values and/or expectations on which
generally hold affectively based attitudes were more their beliefs are based, it has largely increased those
persuaded by a beverage taste test, whereas those appeals’ effectiveness (Cavazza et al., 2010; Lausten,
who generally hold cognitive attitudes were more 2017; Lavine & Snyder, 2000). For example, research
persuaded by facts about the drink (Haddock et al., finds that people with conservative beliefs tend to
2008; Ruiz de Maya & Sicilia, 2004). place greater weight on past achievements and sta-
Moral Bases. Another influential factor in tus maintenance, whereas liberals tend to value
matching is the extent to which people perceive future achievements and status advancement. Thus,
that their attitudes are based on morality (e.g., Lut- when those with conservative (vs. liberal) beliefs
trell, Teeny, & Petty, in press; Skitka & Bauman, receive messages framed in terms of restoring a
2008). For example, consumers might prefer organic desirable past (vs. ensuring a desirable future; Lam-
produce because they believe it is the ethical choice mers & Baldwin, 2018) or in terms of maintaining
(i.e., a moral basis), or because it is a healthier (vs. advancing) status via the purchase of luxury
choice (i.e., a practical basis). When persuasive goods (Kim et al., 2018b), it tends to enhance per-
appeals match these bases, it results in greater per- suasion. Moreover, these political matching effects
suasion (Luttrell & Petty, in press; Luttrell et al., can be augmented in contexts where the individ-
2019). ual’s political beliefs are made salient (Kim et al.,
In addition to an appeal’s broad moral relevance, 2018a) or when the message matches a subtype of
attitudes can also differ in the specific type of one’s political beliefs (e.g., economic vs. social con-
morality underlying them. According to moral foun- servatism) rather than their broader political orien-
dations theory (Graham et al., 2009), an attitude’s tation (Eschert et al., 2017).
moral basis can derive from one of five foundations Knowledge and Value-Expressive Functions. In
(e.g., care/harm and sanctity/degradation). Conse- addition to the variety of bases underlying atti-
quently, even if people share a common moral atti- tudes, there are also variations in the functions that
tude, they can be differentially persuaded by attitudes serve (Katz, 1960). For example, when an
messages targeting different moral foundations. For attitude serves a knowledge function, it helps inform
example, some people might support protecting the people about how they should act toward orPERSONALIZED MATCHING IN PERSUASION 7
evaluate an object, whereas when an attitude serves as the patterned ways in which they think, feel,
a value-expressive function, it helps communicate and behave (i.e., their personality traits) serve as
what is important to them. Considerable research strong targets for matching. Notably, research has
shows that persuasive messages that target the rele- often categorized identities and personality under
vant attitude function tend to be more effective the same umbrella term (i.e., “individual differ-
(Hullett, 2002; LeBoeuf & Simmons, 2010; Shavitt, ences”). Thus, we, too, include them in the same
1990; Spivey et al., 1983; Snyder & DeBono, 1989; section but divide their discussion. We describe
Clary et al., 1994; see Carpenter, 2012). In an illus- identities in terms of the multiple dimensions repre-
trative study, Julka and Marsh (2005) measured the senting individuals’ self-concepts—from those that
extent to which individuals’ positive attitudes are more social to those that are more personal
toward organ donation served a knowledge or a (Oyserman, 2009). We describe personality traits in
value-expressive function. For participants whose terms of their common depiction, namely the Big 5
attitudes served a knowledge function, attitude personality traits (Goldberg, 1990).
change was greater in response to a message that Social Identities. Framing a message as particu-
provided answers to common informational ques- larly beneficial or relevant to a social group to
tions about organ donation rather than a message which the message recipient belongs can enhance
describing the moral support for it. The opposite persuasion. For example, when appeals are framed
was true for participants whose attitudes served a as particularly advantageous for a specific gender,
value-expressive function. those who identify with that gender tend to be
Social-Adjustive Function. In line with the more impacted (Fleming & Petty, 2000; Meyers-
research on attitude bases, attitude functions can be Levy & Sternthal, 1991). Matching messages to con-
attitude-specific or reflected in a consumer’s sumers’ ethnic identities have produced these effects,
broader disposition. The most studied attitude func- too. For example, participants whose Asian identity
tion examined as an individual difference is the so- was made salient responded more favorably to
cial-adjustive function which is assessed with the advertisements targeted toward that identity (Fore-
self-monitoring scale (Snyder, 1974). High self-moni- hand et al., 2002). Matching appeals to consumers’
tors are concerned with their social image, adapting sports fandom (using sports metaphors for those
their attitudes and behavior to fit their current higher on this identity) also increases the message’s
interpersonal circumstance. Low self-monitors pay effectiveness (Ottati et al., 1999). Even matching
little mind to their image and are instead concerned messages to consumers’ financial class (i.e., empha-
about expressing congruence between their internal sizing the capacity for personal control for wealthy
beliefs and outward attitudes and behaviors. Thus, individuals and social relationships for low wealth
higher self-monitoring is associated with having individuals) has produced more positive persuasion
attitudes more based on a social-adjustive function. outcomes (Whillans et al., 2017).
As a consequence, high self-monitors experience a Personal Identities. In addition to social identi-
match when advertisements highlight the social ties, messages can also match personal identities to
image benefits of a purchase (e.g., “this is the drink the same effect. For example, consumers who per-
everyone is talking about”), whereas low self-moni- ceived themselves as more sophisticated or more out-
tors experience a match when advertisements high- doorsy reported greater purchase intentions after
light the performance of the product (e.g., “this is receiving a message that emphasized the appeal’s
the highest quality drink around”). Across adver- relevance for that identity (Summers et al., 2016).
tisements for alcohol (Paek et al., 2012; Snyder & Another example is need for cognition (i.e., the extent
DeBono, 1985), electronics (DeBono & Packer, to which people believe they enjoy thinking;
1991), clothing (Lennon et al., 1988), cars (Zucker- Cacioppo & Petty, 1982), where advertising a pro-
man et al., 1988), and other objects (e.g., Graeff, duct as relevant for those enjoy (vs. don’t enjoy)
1996; Lavine & Snyder, 1996; Shavitt et al., 1992), intensive thinking enhances the appeal for those
when the message content or framing matched higher in this trait (Bakker, 1999; See et al., 2009).
levels of self-monitoring, it typically enhanced per- Other research has observed similar effects for domi-
suasion (see DeBono, 2006, for a review). nance orientation (i.e., assertive messages for those
high in dominance vs. diffident messages for those
low; Moon, 2002), sensation seeking (i.e., unusual
Matching to Identities and Personality Traits
message structures for high sensation seekers vs.
The social and personal ways in which individu- normal message structures for those low; Palm-
als perceive themselves (i.e., their identities) as well green et al., 2002; Self & Findley, 2010), future8 Teeny, Siev, Bri~
nol, and Petty
orientation (i.e., distant advantages/immediate dis- allowed for greater identification of both con-
advantages for those high in future orientation vs. sumers’ identities and personality traits (cf., Barber
a
immediate advantages/distant disadvantages for et al., 2015; Kteily et al., 2019), expanding the possi-
those low; Strathman et al., 1994; Tangari & Smith, bilities for targeting consumers based on these
2012), and many other individual identities (e.g., dimensions. Of course, the employment of such
Coe et al., 2017; Mannetti et al., 2010; Williams-Pie- strategies should be done with ethical considera-
hota et al., 2004, for a review, see Dijkstra, 2008). tions in mind.
Indeed, even self-esteem has been effectively tar-
geted: Advertising nonidealized (vs. idealized) con-
Matching to Cultural Orientations
tent is more effective for consumers who view
themselves unfavorably (Bian & Wang, 2015). As a final category to which messages have been
Personality Traits. The most widely employed matched, we consider the various qualities
approach to categorizing individuals’ personality is bestowed upon people by their cultures. Culture
the Big 5 factors model (Goldberg, 1990), where peo- can shape thinking styles, wherein Western cultures
ple are postulated to differ along five key dimen- tend to emphasize thinking analytically and Eastern
sions (e.g., extraversion and openness). Matching cultures tend to emphasize thinking holistically (Nis-
the content of a message to consumers’ standing bett et al., 2001). Thus, when a message matches a
along one or more of those dimensions has consumer’s culturally shaped thinking style (e.g., a
enhanced persuasion. In an early demonstration product is portrayed in isolation for analytical thin-
(Wheeler et al., 2005), participants who varied in kers vs. as part of a broader context for holistic
their degree of extraversion received a message thinkers), it tends to elicit more positive persuasive
about a video player that was either framed to outcomes (Liang et al., 2011; Monga & John, 2006,
appeal to extraverts (e.g., you’ll be the life of the 2010; Uskul & Oyserman, 2010).
party) or for introverts (e.g., you can enjoy movies Independent and Interdependent Self-Con-
without the crowds). When strong arguments were strual. Perhaps the most studied cultural dimen-
presented, the matched appeal was more effective. sion in personalized matching is a person’s
In another study that looked at the entire Big 5 independent versus interdependent self-construal. People
(Hirsch et al., 2012), individual ads for a cell phone with an independent self-construal view themselves
were developed to match each of the personality as separate and unique from others, whereas those
factors (e.g., the extraversion ad emphasized the with an interdependent self-construal view them-
phone’s ability to help consumers “be where the selves as connected and related to others (Markus &
excitement is,” whereas the neuroticism ad empha- Kitayama, 1991). Although there are individual dif-
sized how it will help them “stay safe and secure”). ferences within any given country (Park et al., 2002)
Regardless of the dimension, when the ads matched and various situational factors (Aaker & Williams,
a person’s more dominant trait, they rated it as 1998; Gardner et al., 1999) that influence the degree
more effective and reported greater intentions to to which a consumer leans more toward indepen-
purchase the phone. Tailoring messages to con- dence versus interdependence, differences on this
sumers’ Big 5 traits was reportedly a prominent aspect typically emerge as a function of one’s coun-
strategy employed to influence U.S. political elec- try of origin. Western consumers tend to be more
tions in 2016 (Hern, 2018). Although some analysts independent, whereas Eastern consumers tend to be
contest the extent of impact this approach had (Gib- more interdependent (Singelis, 1994). Thus, when
ney, 2018), the viability of such a strategy has been Western consumers received a message framed for
confirmed by the laboratory studies just described independent self-construal (e.g., the consequences of
as well as large-scale, digitally implemented field product consumption pertain to the self), and East-
research (Matz et al., 2017). ern consumers received a message framed for inter-
These matching effects for personality traits seem dependent self-construal (e.g., the consequences of
to emerge regardless of whether consumers objec- product consumption pertain to their relationships),
tively versus subjectively possess them (Li, 2016). it was more effective than the mismatched message
For example, whether or not consumers explicitly (Uskul & Oyserman, 2009). Similar findings have
acknowledge themselves as high in neuroticism, occurred when self-construal was either measured or
messages matched to this dimension have increased situationally manipulated for participants within the
persuasion (Hirsch et al., 2012). It is also worth same culture (Sung & Choi, 2011).
mentioning that advances in technology and access Other research has identified more indirect vari-
to consumers’ online and offline activity have ables that can produce similar outcomes. ForPERSONALIZED MATCHING IN PERSUASION 9
example, promotion (vs. prevention)-focused mes- statistical procedures for identifying the “threshold”
sages for independent (vs. interdependent) self-con- where it is optimal to switch participants from one
strual (Aaker & Lee, 2001; Lee et al., 2000; Sherman communication type to another (Joyal-Desmarais
et al., 2011), individual (vs. collective)-focused et al., 2020).
appeals for independent (vs. interdependent) self-
construal (Han & Shavitt, 1994; Zhang & Gelb,
Source-to-Recipient Matches
1996), as well as temporally distant (vs. proximal)
benefits for independent (vs. interdependent) self- The bulk of research on personalized matching
construal (Pounders et al., 2015; Spassova & Lee, has examined the influence of aligning characteris-
2013) have all been shown to produce enhanced tics of the message with those of the recipient.
persuasion, because these kinds of messages are However, personalized matching can also be instan-
more compatible with the person’s culturally tiated by aligning source characteristics—regardless
bestowed self-construal (see Huang & Shen, 2016). of what the message conveys—with that of the
recipient. Broadly, any form of similarity between
the source of the message and the recipient tends to
Practical Considerations for Message-to-Recipient
enhance persuasive impact (e.g., gender, race, and
Matches
motivational orientation; Lu, 2013; Phua, 2014).
Matching a message to aspects of a person relies Although there are fewer studies on source-to-recip-
on the general idea that people can have very dif- ient than message-to-recipient matches, we provide
ferent reasons for liking or disliking things, thereby a few examples below.
leading them to weigh different dimensions in mes- In different lines of research on matching to
sages differently or find particular assertions to be affective and cognitive states, emotionally intense
more compelling than others. In addition to the consumers were shown to be more persuaded by
general categories we have highlighted (e.g., match- more emotionally intense sources (Aune & Kikuchi,
ing to affect versus cognition), the reasons for liking 1993); consumers higher in emotional intelligence
and disliking things can also be quite specific. For were more persuaded by sources matched on this
example, some car purchasers might care primarily dimension (Kidwell et al., 2020); and consumers
about gas mileage whereas others care more about higher in power were more influenced by sources
repairability. Even though both concerns are cogni- who possess a more powerful status (Dubois et al.,
tive and practical rather than affective or moral, 2016; Briñol et al., 2017). Similar findings have
matching the message to the highly specific con- occurred for sources’ motivational orientations,
cerns of the consumer should enhance message where promotion (vs. prevention)-focused con-
effectiveness. Indeed, we presume that the more sumers are more persuaded by sources who exhibit
specific and individualized the message content promotion (vs. prevention)-focused achievement
match is to the person’s concerns, the more effective styles (Lockwood et al., 2002) and/or use more
it can be. However, because this high level of per- eager (vs. vigilant) nonverbal communication styles
sonalization can be impractical, it is useful to know (Cesario & Higgins, 2008).
that even matching to general categories of atti- Still, other work has shown how the source can
tudes, objects and people can enhance persuasion. match to individuals’ relevant attitude functions.
Nonetheless, future research would benefit from a For example, consumers whose attitudes are knowl-
greater delineation of dimensions relevant to a edge- versus social adjustment-based are more per-
specific consumer segment that might reliably serve suaded by sources who serve a matched function:
as targets for personalized matching. expert (knowledgeable) sources for recipients whose
Another practical matter to consider is how the attitudes serve a knowledge function and attractive
consumer’s specific standing on that individual dif- (socially desirable) sources for recipients whose atti-
ference (i.e., higher vs. lower) determines which tudes serve a social-adjustive function (Evans &
kind of matched message would be most effective. Clark, 2012; Ziegler et al., 2005). Moreover, because
For example, consider self-monitoring, where people sources’ identities are often salient for message
low on the scale tend to be influenced by “quality” recipients, this dimension, too, has reliably pro-
appeals and people high on the scale tend to be duced matching effects. For example, recipients
influenced by “image” appeals (Snyder & DeBono, who share the source’s politics (e.g., a liberal source
1989). But, at what point along the scale is it most advocating to a liberal recipient) tend to exhibit
beneficial to switch from a quality to an image mes- greater attitude change (Hartman & Weber, 2009;
sage? Fortunately, recent work has begun to look at Nelson & Garst, 2005), even if the content of the10 Teeny, Siev, Bri~
nol, and Petty
message does not align with their stance (Bochner, personalized matching effects. Work on setting-to-
1996). Additionally, when a source’s culture (which recipient matches has often examined how brick-
can be signaled with as little as an accent) aligns and-mortar atmospherics (e.g., a store’s ambient
with the recipient, this too can enhance persuasion music, scent, and lighting) can interact with con-
(Ivanic et al., 2014; Tsalikis et al., 1992). sumer characteristics to enhance influence. For
One feature of personalized matching that is example, a field study by Morrin and Chebat (2005)
especially applicable to source-to-recipient matches found that an affectively charged induction (i.e.,
(as well as setting-to-recipient matches) is that they emotional music) was more effective at increasing
often occur through physical cues (Guyer et al., in-store purchases for impulsive shoppers, whereas
2019). For example, consumers are more persuaded an induction more linked to contemplation (i.e.,
by sources who physically resemble the consumer’s scent; Chebat & Michon, 2003) was more effective
in-group (e.g., Olivola et al., 2018). Similarly, when for nonimpulsive shoppers.
other physical traits of the source, such as the dom- Similar findings have emerged for consumers’
inance conveyed by their facial appearance (Laust- purchasing goals. Whereas hedonic shoppers (i.e.,
sen & Petersen, 2016) or the pitch of their voice recreational shoppers) report greater in-store pur-
(Banai et al., 2018), align with traits personally chase intentions for highly stimulating stores (i.e.,
matched to the recipient, it can enhance persuasion. with red walls and cluttered floorplans), utilitarian
This is especially true when the source’s physical shoppers (i.e., economic shoppers) report greater
traits match consumers from a minority group, purchase intentions for less stimulating stores (i.e.,
including sources who match ethnic (Brumbaugh, with blue walls and spacious floorplans; van Rom-
2002; Desphande & Stayman, 1994; Whittler & pay et al., 2012). Additionally, consumers’ desire for
DiMeo, 1991) as well as gendered traits (Rosenberg- control can produce similar effects, where con-
Kima et al., 2010). sumers higher in this dimension report greater pur-
chase intentions in noncrowded versus crowded
stores (van Rompay et al., 2008).
Setting-to-Recipient Matches
With online shopping becoming an increasingly
As a final form of personalized matching, mar- dominant venue for consumer behavior, marketers
keters can match an aspect of the setting or context can also utilize the atmospherics of the digital envi-
in which persuasion occurs (i.e., variables in the ronment to create personally matched appeals (for a
consumer’s environment; Mehrabian & Russell, review, see Wan et al., 2009). For example, whereas
1974) to an aspect of the message recipient. Setting- hedonic online shoppers respond more favorably to
to-recipient matches could entail a characteristic of immersive and experiential websites, utilitarian
the setting aligning with an already established online shoppers find these features distracting (Hun-
characteristic of the recipient (e.g., German versus ter & Mukerji, 2011). In contrast, utilitarian shoppers
French consumers exhibiting different purchase respond more favorably to task-oriented features,
intentions as a function of playing German versus such as the website’s navigability and search func-
French music), or the setting can activate a personal tions (Gounaris et al., 2010; see also Bridges & Flor-
characteristic that then matches the setting. For sheim, 2008). Other research demonstrates the value
example, in one study (North et al., 1999), playing of considering less controllable atmospherics in these
stereotypically German (vs. French) music in a store online settings, such as the time of day when a per-
setting activated a greater self-perception of Ger- suasive message is read. For example, consumers
man (vs. French) working knowledge in customers, categorized as morning types exhibited greater atti-
thereby increasing the sale of German (vs. French) tude change following a message with strong argu-
wine. ments delivered during the morning (vs. the
In general, the personalized matching literature evening; Martin & Marrington, 2005).
has conducted less research on setting-to-recipient
matches, potentially due to the relative difficulty in
The Meaning of the Match
personalizing the consumers’ environment or the
reduced ability to target a specific segment as the We have now covered a wide range of dimen-
setting applies to all consumers who enter it. sions upon which personalized matching has
Nonetheless, if an aspect of the consumer’s setting enhanced persuasiveness. As mentioned already,
matches a consumer characteristic (e.g., delivering a however, this is not always the case. Sometimes,
message in a quieter room for introverts vs. a lou- personally matched messages can reduce persua-
der room for extraverts), it can produce sion. In order to understand why and when thisPERSONALIZED MATCHING IN PERSUASION 11
occurs, it is useful to consider the valenced meaning personalized matching can produce. Nonetheless,
that the match generates for consumers (i.e., and perhaps most commonly, a personalized match
whether the matching in a message is interpreted can engender a negative meaning if it is interpreted
as positive or negative). For example, two people as an invasion of privacy (e.g., van Doorn & Hoek-
might each perceive a matched message to be famil- stra, 2013; White et al., 2008). Other research has
iar, but that familiarity could be interpreted posi- shown that matched messages can generate a nega-
tively by one person (e.g., desirable because of its tive meaning if they are interpreted as an attempt at
fluency) or negatively by another (e.g., undesirable manipulation (Bri~
nol et al., 2015; David et al., 2012;
because the content seems already known). Regard- Reinhart et al., 2007), or if they are interpreted to
less of how the meaning emerges (either implicitly be based on an unfair or stereotypic judgment about
or explicitly) or the specific experience that gener- the person (Derricks & Earl, 2019; Kim et al., 2019a;
ated the meaning, the positivity or negativity of White & Argo, 2009). Additionally, a personalized
that meaning is a key factor in determining the match can elicit a negative meaning if the person
effect of the matched content (Bri~ nol et al., 2018). interprets the message as containing already known
However, as we will see shortly, positive meanings content (Clark et al., 2008; Hastie, 1984; Mah-
do not invariably lead to positive persuasion out- eswaran & Chaiken, 1991). Although there is less
comes and the converse is also true for negative research on the emergence of negative meanings,
meanings. they are nevertheless important to consider.
Positive Meanings Eliciting Positive versus Negative Meanings
In general, personalized matching has been Understanding whether a personalized match
shown to produce positive meanings in most cases generates a positive or negative meaning is critical
and this can stem from several sources. Perhaps for understanding how it will affect persuasion.
most obviously, personalized matches can produce Although people predominately appear to interpret
positive meanings due to the perceived self-relevance personalized matches positively, certain individual
of the match (DeBono & Packer, 1991; Petty & differences and situational factors might be espe-
Wegener, 1998; Abrahamse et al., 2007). Because cially likely to undermine this. For example, people
most people have positive feelings toward them- might be more likely to interpret a personalized
selves, anything linked to the self can also take on match negatively if they hold an interdependent
that positivity (Gawronski et al., 2007; Horcajo et al., (vs. independent) self-construal, because interde-
2010a). Indeed, recent neuroscience research demon- pendent consumers prefer to think about them-
strates that matched messages activate individuals’ selves as part of the group (vs. individuated;
ventromedial prefrontal cortex, which is associated Kramer et al., 2007; Li et al., 2011). Additionally, if
with self-relevant evaluations (Aquino et al., 2020), consumers are marketing savvy (Friestad & Wright,
thereby potentially increasing favorability toward 1994), they might be more likely to interpret a per-
matched appeals. Nonetheless, other positive mean- sonalized match negatively (e.g., as a manipulation
ings that matched messages can generate include a attempt) versus positively (e.g., as meaningful to
greater feeling of rightness or fit (Cesario et al., 2004; me; Maslowska et al., 2013). Other research sug-
Higgins, 2005; see Cesario et al., 2008), familiarity or gests that consumers’ concerns about data privacy
fluency (Labroo & Lee, 2006; Thompson & Hamilton, could also affect their reaction to personalized mes-
2006), self-efficacy in attaining the advocated outcome sages, where consumers higher in data privacy con-
(Han et al., 2016; Sherman et al., 2006; Bostrom cerns are more likely to interpret a personalized
et al., 2013), and authenticity (Bleidorn et al., 2016; match negatively (i.e., as an invasion of privacy;
Harms et al., 2006). Although the present research Hoffman et al., in press). Negative meanings are
has treated the consequences of different positive also likely if the message is too highly matched
meanings as relatively interchangeable, future (e.g., it includes a consumer’s prior transaction his-
research would benefit from determining whether tory; van Doorn & Hoekstra, 2013), or if the con-
specific positive meanings have specific effects. sumer recognizes that the information for matching
the message was collected on a website separate
from where the message is being delivered (Kim
Negative Meanings
et al., 2019a). Across these examples, one common
As noted, there has been significantly less work theme is that if consumers become explicitly aware
identifying the potential negative meanings that of the personalized match (see Cesario et al., 2004),12 Teeny, Siev, Bri~
nol, and Petty
then inferences about the source of the personalized much of this research has been unified and inte-
information or the motives behind it are more likely grated under the ELM (Petty & Bri~ nol, 2012; Petty
to be negative. & Wegener, 1998). In the ELM, different mecha-
Further research is needed to determine which nisms are categorized into a finite number of gen-
factors lead a consumer to become aware of and eral attitude change processes which emerge as a
interpret such personalization negatively. For exam- function of individuals’ levels of elaboration. That is,
ple, when might a feeling of familiarity from a in any given persuasion context, consumers can
match instill a positive reaction (e.g., a feeling of range from relatively low to high in how much
comfort) versus a negative reaction (e.g., a feeling they think about and scrutinize the message.
of boredom)? Some research has suggested that Depending on where the person falls along this
consumers are less likely to generate negative pri- elaboration continuum, personalized matching can
vacy concerns regarding personalized messages if influence consumers’ attitudes and behaviors
they have previously consented to allow access to through different types of psychological processes.
their data (Hoffman et al., in press). Or, what if the The ELM holds that it is critical to understand the
match simultaneously generates a positive and neg- type of process through which the match has its
ative meaning? In any case, a better understanding effects in order to elucidate whether it will increase
of when matches generate different meanings is or decrease persuasion as well as whether those ini-
important, because as argued shortly, these mean- tial results are likely to be durable and impactful
ings can play a critical role in determining the (Petty et al., 1995).
match’s eventual effect on persuasion. Individuals’ degree of elaboration is determined
by their motivation and ability to think critically
about the message. The motivation to process a
message can be affected by situational factors, such
Multiple Processes in Personalized Matching
as the likelihood that a consumer is in the market
So far, we have shown there are many ways in for and will be able to purchase a product (Petty
which one can use the message, source, or setting et al., 1983), or dispositional factors, like their gen-
to create a match with at least one of the aspects of eral preference to engage in careful thinking
the recipient. And from our discussion so far, it (Cacioppo et al., 1983). Similarly, one’s ability to
might sound like matches generating positive process a message can be affected by situational
meanings invariably produce positive outcomes, factors such as the speed at which the message is
whereas negative meanings produce the reverse. delivered (Smith & Schaffer, 1995) or the number of
Unfortunately, this analysis is too simplistic. To times it is repeated (Cacioppo & Petty, 1989) as
understand how positive meanings can sometimes well as dispositional factors like a person’s intelli-
produce negative outcomes and negative meanings gence (Wood et al., 1995). Depending on a con-
can sometimes produce positive outcomes, we turn sumer’s situational or dispositional state of
to an analysis of the multiple mechanisms by which elaboration, the ELM holds that the process by
personalized matching can impact attitudes. which a personalized match affects persuasion can
Although there have been periodic calls for match- change (Petty et al., 2000).
ing researchers to pay more attention to underlying More specifically, if consumers are constrained to
processes (e.g., Hawkins et al., 2008), prior reviews be rather low in their thinking (e.g., many distrac-
have tended to focus primarily on matching effects tions are present), personalized matching is
and outcomes, as we did above. To elucidate the expected to function as a simple cue or heuristic in
mechanisms of matching, we rely on a well-estab- the attitude change process. In these instances, the
lished general theory of persuasion, the elaboration match produces an effect consistent with its
likelihood model (ELM; Petty & Cacioppo, 1986; refer valenced meaning. For example, a person might
to Figure 1 for an overview of the ELM applied to quickly reason that “I like this because it’s similar
matching effects). After a brief review of the ELM, to me,” or “I don’t like it because it is trying to
we show how it can be useful in explicating when manipulate me,” without much thought about the
matching enhances persuasion and when it can merits of the advocacy. In contrast, if elaboration is
backfire. constrained to be high (e.g., people are incentivized
Decades of research have striven to describe and to think about the evidence presented), matching
model the fundamental underlying processes of can influence persuasion through more effortful
attitude change (for reviews, see Bri~ nol & Petty, processes. For example, the personalized match can
2012; Eagly & Chaiken, 1993; Maio et al., 2019), and be carefully scrutinized as an argument in and ofYou can also read
