Recently, manipulative techniques, such as dark patterns, are widely applied. However, there is a need for clarification regarding these techniques and related phenomena. In particular, there is still no clarity about the terminology and conceptual basis of consumer manipulation. This paper aims to address this shortcoming by introducing a definition and classification of consumer manipulation.
This paper takes a conceptual approach, drawing on existing literature and established theories to comprehend the phenomenon of consumer manipulation.
The paper proposes that consumer manipulation comprises three dimensions: limited transparency, perceived restriction of autonomy and the feeling of being tricked. This paper presents a classification of different types of consumer manipulation and demonstrates how these types can lead to varying outcomes based on the source of manipulation, changeability, reasons for perception and perception timing. Based on this classification, research propositions informed by established theories and concepts are presented.
To the best of the author’s knowledge, this work represents a first attempt to address the concept of consumer manipulation through a clear definition and classification. It is relevant for academics, policymakers and practitioners since it facilitates informed discussions and analyses of this concept. It benefits companies by raising awareness of potentially overlooked consequences of manipulative tactics, while also benefiting consumers by reducing their exposure to manipulation.
Introduction
One goal of consumer policy is to create a safe digital space, in which consumers’ and their rights are protected (European Commission, 2021). This is particularly important in digital world’s day-to-day reality, where unethical behavior from organizations impacts consumer decision-making. There, beside others, the use of manipulative techniques, for example, in the form of exploiting human biases and heuristics, as well as new technologies for consumer manipulation, is common for many companies (Luguri and Strahilevitz, 2021; Susser et al., 2018). To name a few, manipulative techniques, such as the manipulation of reviews (Zhuang et al., 2018), behavioral targeting (Murphy et al., 2013) and drip pricing (Rasch et al., 2020), are widely used in the digital world, especially in e-commerce (e.g. Voigt et al., 2021) and gaming (e.g. Søraker, 2016). As a consequence the term “dark patterns” has recently gained popularity (e.g. Cara, 2019; Di Geronimo et al., 2020; Mathur et al., 2019). As Dark Patterns are primarily applied to increase shareholder instead of consumer value (Gray et al., 2018) the phenomenon seems unethical (Sheth and Sisodia, 2007).
Facing these developments, (consumer) policy makers and authorities initiated some interventions. The European Commission recently passed the Digital Services Act and the Digital Markets Act which among others aims at prohibiting the use of dark patterns (European Parliament, 2022a, 2022b; BEUC, 2022). Even though the use of these techniques is highly questionable from an ethical perspective (e.g. Gray et al., 2018), it is also questionable if the use of manipulative practices is beneficial from a business perspective.
From a company’s perspective, consumer manipulation has – on first sight – primarily positive effects. It has been proven that manipulation – e.g. in form of dark patterns (Susser et al., 2018) or sludges (Thaler, 2018) – leads to higher subscription rates (Luguri and Strahilevitz, 2021; Thaler, 2018). Nevertheless, empirical research indicates that – in the long run – this might back-fire (see e.g. Waltenrath, 2021; Voigt et al., 2021.). A prominent example for this is the Cambridge Analytica scandal in 2016, in which the social network Facebook manipulated their users during presidential elections in the USA. As a result, 44% of US-Facebook-users between 18 and 29 years deleted their accounts after the manipulation became public (LaMagna and Passy, 2018; Perrin, 2018). However, consumer manipulation occurs in e-commerce as well. Consider for example buying flowers online and later on realizing that a chargeable greeting card was added to your cart without notification (Lukoff et al., 2021). However, despite initial and rare evidence from other research areas (e.g. legal; Susser et al., 2018) about potential negative effects, to date, little is known about consumer’s reactions to manipulations and the underlying mechanisms (Moran, 2020).
It is unclear what is meant by the term “consumer manipulation” and which theoretical mechanisms might drive its negative effects. This paper aims to clarify the concept of “consumer manipulation” and is structured as follows: At first a definition of consumer manipulation in the digital world will be presented, to deliver clear terminology and conceptual understanding. This includes a differentiation of consumer manipulation from related terms, such as persuasion or nudging. Then a classification of consumer manipulation will be suggested, which is followed by avenues for further research in form of several propositions and the derivation of some implications for research as well as management.
By doing so, this paper adds to the existing body of literature of consumer manipulation and dark patterns. Especially, this paper contributes to this research stream by clarifying and conceptualizing consumer manipulation. Hopefully, this will enable scholars and further research to better examine this phenomenon.
Literature review
To date, there are very few articles in the literature dedicated to the topic of consumer manipulation. Only Strang et al. (2015) and Li et al. (2021) contribute to the definition of the term. Strang et al. (2015) define the term as “to control or play upon by artful, unfair or insidious means, especially to one’s own advantage” (Strang et al., 2015, p. 248). A more implicit description of the term is provided by Li et al. (2021), who describe manipulation as “users’ concern that a social media platform may intentionally mislead them through means such as providing vague or partial information” (Li et al., 2021, p. 7). Other articles use the concept of consumer manipulation to some extent, but do not directly define the term. For example, Hietanen and Rajamäki (2018) classify subscription traps as a form of customer manipulation.
From a theoretical point of view, the Persuasion Knowledge Model (Friestad and Wright, 1994) and the Information Manipulation Theory (McCornack, 1992) in particular could be used to explain consumer manipulation. While the persuasion knowledge model could explain the reactions to any manipulation, the information manipulation theory seems suitable for explaining the emergence of manipulation.
The Persuasion Knowledge Model (Friestad and Wright, 1994) aims to explain how people build up and use knowledge about persuasion attempts. According to Friestad and Wright (1994), individuals develop persuasion knowledge over time, which enables them to recognize manipulations and counter them with coping strategies. Information Manipulation Theory (McCornack, 1992) aims to explain what makes messages manipulative. According to McCornack (1992), messages have a manipulative effect on the recipient if they violate at least one of four principles of communication (McCornack, 1992). These principles are based on Grice’s four communication maxims: the quantity maxim (appropriate amount of information), quality maxim (truthfulness), relevance maxim and clarity maxim (Grice, 1989).
Conceptualization of consumer manipulation in the digital world
Definition of consumer manipulation
Since there is still no consensus on a clear definition of consumer manipulation in the digital world, definitions of manipulation from other disciplines might be useful to arrive at a clear characterization. Considering the term’s etymology, the term “manipulation” is derived from the Latin term manipulus (a handful), which in turn is composed of manus (hand) and plere (fill) (Pfeifer et al., 1993). The term was used as a unit of measurement in pharmacy in the 18th century. In later years, the term manipulation was used as “skillful handling” or “artifice” and since the 19th century has primarily described the handling or management of people (Online Etymology Dictionary, 2023; Pfeifer et al., 1993). Also today, in general, the term “manipulation” describes an attempt to influence a person’s decision, whereby the manipulator focuses on his own interests without regarding the interests of the manipulated person (Sunstein, 2015). According to Sunstein (2015), a person’s or organization’s action can be considered as manipulative “to the extent that it does not sufficiently engage or appeal to people’s capacity for reflective and deliberative choice” (Sunstein, 2015, p. 239). Thus, manipulations make consumers feel as if they could not fully make a decision on their own (Sunstein, 2015). Therefore, manipulation can be described as “deceptively influenc[ing]” (Goodin, 1980, p. 19) someone. Rudinow (1978) explains manipulation as “the attempt to motivate another’s behavior by altering her complex of goals by means of deception or by playing on a supposed weakness she has” (Rudinow, 1978, p. 346). Sher (2011) draws on this definition and adjusts it by defining manipulation as “undermining what the agent believes is his/her audience’s normal decision-making process by means of deception or by playing on a supposed weakness she has” (Sher, 2011, p. 99). The author points out that manipulation is not about increasing a person’s original motivation but changing the behavior in favor of the manipulating party by undermining the normal decision-making process (Sher, 2011). This limits a person’s ability to make a self-determined decision (Sher, 2011). Complementary, Baron (2003) distinguishes between three types of manipulation:
Deception. which describes the use of lies, false promises or misleading by giving false information (Baron, 2003; Wood, 2014). For example, if a car dealer states that the car for sale has never been involved in an accident, thereby exploiting an information asymmetry, this would be deception.
Pressure to acquiesce. means aiming to reduce the resistance to the manipulative action. If the car dealer, for example, asks the potential customer to decide within the next two hours, otherwise he has wasted his working time, it would be pressure to acquiesce.
Playing upon emotions, emotional needs or character weaknesses whereby for example “vanity and the need for approval” as well as greed might be used for manipulation (Baron, 2003; Wood, 2014). Finally, if the car dealer tells the potential customer that he can understand if the potential customer has overestimated his financial situation and cannot afford the car after all, he would be alluding to vanity and thus playing upon emotions, emotional needs or character weaknesses would be present.
In addition, drawing on Goodin’s (1980) thoughts, Handelman (2009) specifies trickery as a central aspect of manipulation. Trickery is described as the central factor that distinguishes manipulation from, among other things, clear deception (Handelman, 2009; Noggle, 2018).
It should be noted that manipulation is not a binary construct but a continuum (Sunstein, 2015). Susser and colleagues (2018) describe hiddenness, exploitation of decision-making vulnerabilities and targeting as central aspects. While hiddenness is a necessary characteristic, the others are not mandatory but core to a successful manipulation (Susser et al., 2018). Therefore, hiddenness is an important characteristic, because if a person is aware of being influenced or manipulated, this could be taken into account when making a decision and thus the decision-making process would not be undermined (Baron, 2003; Susser et al., 2018).
In the context of consumer decision making, selectively presenting information by a company as well as using false or exaggerated claims are three main strategies that can be classified as manipulation (Gatignon and Le Nagard, 2016; Greenspan, 2003). Thus, the presentation of selective information refers to communicating only positive aspects of a product or service and hiding less favorable attributes (Gatignon and Le Nagard, 2016). This is accompanied by an exaggeration of some information as an extremely positive presentation of individual aspects. Appealing to emotions also refers to a tactic, aimed at evoking consumers’ emotions (Gatignon and Le Nagard, 2016).
Another important feature of manipulation is transparency. Regarding manipulation, transparency means being nontransparent about the manipulation itself or hiding certain characteristics of an information to influence a person’s decision (Sunstein, 2015). Thus, manipulation in its nature is most effective when it occurs nontransparent and is therefore typically hidden (Baron, 2003).
Although the presented definitions may differ with regard to individual aspects, they consider similar components. According to them, manipulation is used to motivate a person to show a targeted behavior (Handelman, 2009; Rudinow, 1978; Sher, 2011) that they might not have shown without the manipulation (Susser et al., 2018). To achieve this, manipulation typically uses deceptive elements, a person’s weakness, emotions or emotional needs (Baron, 2003; Goodin, 1980; Rudinow, 1978; Sher, 2011). As a result, manipulation undermines the normal decision-making process of a person in a hidden way (Sher, 2011; Susser et al., 2018) and does not sufficiently include the person’s capacity for reflection and deliberation or interfere with the target person’s mental autonomy without restricting her physical autonomy (Handelman, 2009; Sunstein, 2015). This can be specified in that the consumer in the digital world might have the illusion of making a decision completely by himself and being able to decide freely (Handelman, 2009; Sher, 2011).
Summarizing relevant aspects of existing definitions of manipulation, consumer manipulation in the digital world might be defined as:
[…] any action aiming to influence the decision-making process of a consumer in the digital world in a hidden, targeted and often selfish manner with the understanding that the consumer has the feeling that he or she is free to make a decision, even though a large part of the decision was determined in advance by the organization’s activities.
Dimensions of consumer manipulation in the digital world
As the definition of consumer manipulation in the digital world suggests, there are multiple dimensions of the construct, which can be summarized to three main components.
Consumer manipulation:
refers to hidden tactics, which undermine or restrict the consumer’s autonomy (Susser et al., 2018);
is characterized by limited transparency and bad information quality (Goodin, 1980; McCornack, 1992); and
uses trickery and thus, creates a feeling of being tricked (Handelman, 2009; Sunstein, 2021).
A manipulated person cannot endorse herself because her capacity for deliberation is undermined (Sunstein, 2021; Susser et al., 2018). Manipulation strongly restricts a person’s freedom of choice, which can be defined as the “ability to make and enact decisions on their own, free from external influences imposed by other agents” (Wertenbroch et al., 2020, p. 430).
Wertenbroch and colleagues (2020) differentiate actual and perceived autonomy. While actual autonomy describes the extent to which a person can decide on their own; perceived autonomy reflects the person’s evaluation of being able to decide on their own (Wertenbroch et al., 2020). Recommendation algorithms, for example, can increase perceived autonomy but decrease actual autonomy simultaneously (Wertenbroch et al., 2020).
Another example is the dark pattern of “hidden subscriptions” which refers to services or subscriptions that automatically renew and charge the customer recurring fees (Mathur et al., 2019). At first glance decision-making autonomy over the subscription lies with the customers; although autonomy is highly restricted due to the subtle automatic renewals.
With regard to the second dimension - bad information quality - McCornack (1992) argues that a message can be divided into four aspects that could be used for manipulation:
the quantity of relevant information;
the quality of the information;
the manner of the presentation of information; and
the relevance of conversational attributions.
In general, the recipient of a message expects these four criteria to be appropriate. However, if these expectations are not met, it can be considered as manipulation (McCornack, 1992). For example, if an online shop automatically filters out all negative product reviews, this could violate both the relevance and quantity maxims and should be considered manipulation (Xiao and Benbasat, 2011).
Regarding the third dimension, the feeling of being tricked also plays an important role in consumer manipulation in the digital world (Handelman, 2009; Sunstein, 2021). This feeling arises when the manipulator undermines the normal decision-making process, creating a flawed mental state (Noggle, 2018). This might be demonstrated by the following example: If a customer is informed on a hotel booking platform that only one room is still available, but after booking, the room is still available, this is likely to cause the feeling of being tricked.
In this view, it can be stated that perceived consumer manipulation includes the three dimensions of restricted autonomy, bad information quality as well as the feeling of being tricked.
Differentiation of consumer manipulation from related terms
Although manipulation shares some characteristics with similar constructs such as coercion, deception and persuasion, there are still clear differences (Handelman, 2009; Greenspan, 2003). For a sound definition, below, four constructs will be distinguished.
Deception.
Several authors define deception as a subset of manipulation (Baron, 2003; Sher, 2011; Susser et al., 2018). While manipulation resembles deception in terms of feeling betrayed and outraged (Sunstein, 2015), Sunstein (2021, p. 3) distinguishes both by stating that “manipulators trick you; deceivers lie to you.” Additionally Handelman (2009) differentiates both concepts by stating that if false information or unfulfillable desires are used, the action is to be classified as deception instead of manipulation, since manipulation generally gives incentives and does not directly mislead a person (Handelman, 2009). However, the author also explains that “false information” can be used in manipulation, but they are hidden in an elusive language, symbols or hints. In addition, manipulation usually includes hiding certain information (Handelman, 2009). Following this, according to Sher (2011, p. 104), an action is to be considered as deceptive if “it intends to bring about consumer misconception by providing what the marketer believes is false evidence, omitting key evidence, or misrepresenting what the evidence means.”
However, Susser and colleagues (2018) argue that manipulation can be deceptive, but does not have to. In addition, with regard to the definition of manipulation, which refers to undermining the normal decision-making process (Sher, 2011), deception should be considered as an instrument of manipulation, contrary to Sunstein’s appraisal (Sunstein, 2015). Due to the deception, people might become unaware of their options, therefore might becoming unable to decide as they want to Susser et al. (2018). Consequently, deception can be considered as an extreme form of manipulation that consciously misleads a person (Handelman, 2009; Sher, 2011; Susser et al., 2018). In conclusion, deception can be classified as manipulation in most cases, but manipulation is not always deception. The distinction shall be clarified by the following example: If a hotel booking platform states that only one room is available, this would be manipulation if the limited availability is true. However, if the platform uses lies or false information, it is deception.
Coercion.
What manipulation differentiates from coercion is that when a person is manipulated, the person itself finally decides for an option (Wood, 2014), which is not the case with coercion. Thus, manipulation does not make an option unacceptable, but less attractive (Wood, 2014). Another approach to differentiate coercion from manipulation might be that the former limits someone’s choices in a physical manner, while manipulation only limits choices artificially (Handelman, 2009). Thus, coercion would lead to a restriction or absence of freedom of choice, while manipulation suggests an illusionary freedom of choice (Handelman, 2009). While coercion usually makes use of irresistible incentives and somehow forces a person to behave in a certain way (Susser et al., 2018), incentives used in manipulation are typically resistible (Rudinow, 1978). Colloquially, both concepts can be distinguished as follows: “when a person is coerced that person feels used, when a person is manipulated that person feels played” (Susser et al., 2018, p. 17). Overall, while manipulation makes an option less attractive (Wood, 2014), coercion physically restricts choices, forcing a certain behavior (Handelman, 2009). For example, if an online platform makes it difficult to cancel a subscription, this could constitute manipulation. If, on the other hand, cancellation is made physically impossible, for example by a faulty cancellation website, the consumer would be forced to remain in the relationship. The latter would be a case of coercion.
Persuasion.
Persuasion can be characterized as an intervention trying to lead a person in a specific direction regarding a desired choice without restricting existing options (Handelman, 2009). Generally, persuasion might refer to any intervention to change a person’s mind and is therefore a similar concept to manipulation. However, unlike manipulation, persuasion attempts to achieve this change of mind by using rational arguments or incentives and encouraging reflection and evaluation (Susser et al., 2018). This influences the perception of different options (one could also refer to this as convincing), but freedom of choice actually exists, whereas in manipulation it is only an illusion (Handelman, 2009), since options are artificially restricted (Greenspan, 2003). This is one key difference between the constructs. Furthermore, manipulation also differs from persuasion, since persuasive practices usually focus on facts as well as reasons and are neutral and fair (Sunstein, 2015). Thus, persuasion might be manipulative, but not every attempt to persuade is manipulation (Sher, 2011). A persuasion attempt should be considered as manipulative when the ability to decide rationally is undermined by manipulation or deception (Baker and Martinson, 2001).
Consequently, manipulation can be located on a continuum between rational persuasion and coercion. What coercion and persuasion have in common is that both attempts leave the power of decision-making to the person whose behavior is to be influenced. However, they differ because persuasion leaves all options available, while coercion leads to only one option being acceptable (Susser et al., 2018). Thus, with regard to manipulation, a general difference between coercion and persuasion on the one hand and manipulation on the other hand is, that while coercion and persuasion appeal to a person’s evaluation of options, manipulation tries to undermine the whole decision-making process. In an online context, providing consumers with reviews might be persuasive. However, as soon as, for example, only positive ratings are displayed or ratings themselves are created, it would be more of a manipulation.
Nudging.
A nudge describes “any aspect of the choice architecture that alters people’s behavior in a predictable way without forbidding any options or significantly changing their economic incentives” (Thaler and Sunstein, 2009, p. 6). Nudging targets the choice architecture of a person (Susser et al., 2018; Thaler, 2018). That said, nudging as well as consumer manipulation try to influence a consumer’s behavior without force (Susser et al., 2018). Moreover, a nudge could, in principal, be considered as manipulative “to the extent that it does not sufficiently engage or appeal to their capacity for reflection and deliberation” (Sunstein, 2015, p. 239). However, according to Sunstein (2015), most nudges are educational and sufficiently appeal to the capacity of deliberation, which is why they should not be considered as manipulative (Sunstein, 2021). Nevertheless, some nudges - especially applied in the private sector - can be manipulative as well (Sunstein, 2021). This corresponds to the argumentation of Susser and colleagues (2018), who suggest distinguishing between manipulative and non-manipulative nudges using hiddenness as a differentiating aspect. When a nudge is used to influence someone’s decision-making process in a hidden way, it can be considered as manipulation (Susser et al., 2018). For example, if the purpose of a nudge is not hidden, this is not to be considered as manipulative (Sunstein, 2015; Susser et al., 2018). Consequently, nudges are generally transparent and try to fix vulnerabilities, while manipulation tries to exploit them. Nevertheless, the intention and the setting of the nudge should also be taken into account. So if an influence of a consumer’s choice architecture is applied in the private sector and aims to increase not welfare, but a firm’s profit, this is referred to as a “sludge” (Sunstein, 2018; Thaler, 2018) and is considered to be manipulative.
In summary, hiddenness, perceived autonomy and restriction of choice are central to distinguishing manipulation from related constructs. Figure 1 presents the conceptual relationship between these constructs.
Conceptual relation between manipulation, coercion, deception, nudging and persuasion
Conceptual relation between manipulation, coercion, deception, nudging and persuasion
Classification of consumer manipulation in the digital world
For a classification of perceived consumer manipulation in the digital world, the source of manipulation should first be differentiated into direct and indirect consumer manipulation. While direct consumer manipulation refers to organizations that directly manipulate consumers by themselves, indirect consumer manipulation refers to “peer-to-peer” manipulation, which means that one consumer manipulates another. Nevertheless, the manipulating consumer is usually motivated by an organization, meaning that in both cases the manipulation is initiated by an organization. Consequently, the first dimension of the classification is the directness of manipulation, including the characteristics direct and indirect. Second, the irreversibility of a manipulation is also characterizing for perceived consumer manipulation in the digital world. Therefore, the dimension changeability with the characteristics high and low is included in the typology. Furthermore, it is to differentiate whether a manipulation is perceived by the consumer without any external influence or by a third party, e.g. an organization or another consumer. Thus, the next dimension is the reason of perception characterized as self or external. Consumer manipulation in the digital world can also be differentiated by perception timing. Taking the stages of customer experience (Lemon and Verhoef, 2016) into account, it can be distinguished between the perception of manipulation before (ex ante) or after the purchase (ex post). Consequently, the fourth dimension is represented by the time consumers perceive the manipulation, including the characteristics ex ante and ex post.
The interference of different combinations of characteristics of the described dimensions allows a classification of different possible consequences or outcomes (e.g. anger and loyalty) of the applied manipulations. Additionally, different combinations of characteristics lead to different levels of perceived consumer manipulation in the digital world.
In general, it can be assumed that the earlier the manipulation is noticed by the consumer the lower is the anger of the manipulated person (Kim and Li, 2009). This might be because in an early stage of the purchase process, lower costs (e.g. search costs/time) occurred. With this in mind, one can assume that in a later stage of the purchase process the manipulated consumer has invested more time (or psychological costs) thus feeling angrier, compared to a perception in an early stage, because overall transaction costs increase. This corresponds with previous findings that high transaction costs negatively influence customer satisfaction and loyalty (Kim and Li, 2009). In contrast, early perceptions of consumer manipulation can lead to ending the purchase process or using of persuasion knowledge, as demonstrated by Campbell and Kirmani (2000). This coping behavior might immediately impact the financial performance of an organization. As an example, Warren et al. (2021) found that asking for a tip before completing service in a restaurant is considered as manipulative and therefore negatively impacts the tip intention (Warren et al., 2021). Regarding the digital world, this also applies to freemium products, for example, as users state that they felt manipulated when they were asked to pay for a full version of a digital service (Gray et al., 2021).
Considering the initial reason of perception, it can be assumed, that an own perception leads to less anger and less perceived strength of manipulation. This is based on the assumption that consumers believe that the manipulation is not as bad as it could be if they are able to identify the manipulation by themselves. With regard to manipulations that are perceived because of third parties (e.g. other consumers), it can be assumed that this has a greater impact on negative emotions, because consumers will assess the manipulation as unfair and hidden, so that they did not have the chance to notice it by themselves.
Furthermore, based on the reactance theory (Brehm, 1966), it can be assumed that the level of changeability of a manipulation also has an impact on negative affect and emotions. If a manipulation is easy to avoid (e.g. default settings), the negative outcomes should be relatively low. Likewise, the (perceived) autonomy is highly restricted when changeability is low. Therefore, it can be assumed that low changeability leads to higher negative emotions and psychological reactance.
Depending on the source of manipulation, different effects on attitudes towards the manipulation are imaginable. If the manipulation is carried out by a third party (e.g. an influencer), negative effects on attitudes should be lower compared to the situation, when the manipulation was carried out by the organization itself. This assumption is based on attribution theory (Bitner, 1990; Folkes, 1984). If a third party carries out the manipulation, the inconveniences perceived by the consumer might be attributed to the third party. Otherwise, this would be attributed to manipulator (i.e., the organization), if it applies the manipulation itself. Especially manipulated online-reviews and ratings, which become increasingly widespread in the digital world, can be assigned to this category. However, since consumer manipulation in the digital world, is characterized by a targeted and often selfish manner, not all online-reviews can be considered as manipulative. Only those that are written in a selfish and targeted manner are considerable. This would be the case for example, if the organization pays their consumers to write a review.
Summarizing these considerations, a direct manipulation with low changeability that is perceived ex post internal noticed would lead to the maximum negative outcomes (e.g. anger) from an organization’s perspective. On the other hand, the most negative effects would occur if the manipulation is applied (directly) by the manipulator (i.e. the organization) and is perceived ex post due to a hint from another person and is characterized by low degree of changeability. The consequences of all other combinations lie between these two.
Based on the classification presented above, all manipulation attempts can be classified. For example, consider the dark pattern typology introduced by Gray et al. (2018): The dark pattern nagging interrupts ordinary consumer interaction to trigger an activity that is more beneficial to the company. First, this is a direct manipulation, as it is carried out by the organization itself. Second, the changeability is low because the consumer cannot do anything about the interruptions. Finally, the consumer will most likely perceive the manipulation himself before a transaction (ex ante). According to this classification, nagging would be a direct manipulation with low changeability that is perceived by the consumer himself before the purchase.
In Figure 2, all possible combinations of the dimensions are presented.
Future research
The previous explanations provide manifold propositions and suggestions for future research on consumer manipulation. First, future research needs to empirically validate the presented classification. As argued before, it can be proposed that:
The consequences of perceived consumer manipulation depend on the source of the manipulation, the changeability, the reason for perception and the perception timing.
Further, there are several contextual and individual factors that might be important when examining consequences of consumer manipulation and their intensities. From this, further propositions for future research, focusing on key performance indicators for customer relationships (e.g. satisfaction, loyalty), as these are particularly important for a company’s success and seem most at risk from consumer manipulation, will be derived.
Further, as argued above, perceived consumer manipulation is based on artificial restriction of the consumer’s autonomy, bad information quality and the feeling of being tricked. These factors can occur either ex ante or ex post to the decision process and lead to different short-term as well as long-term effects. If the consumer notices the manipulation before buying a product or service, he will probably stop the decision making process. In the case that the consumer does not notice the manipulation before, but after buying the product or service, the company may generate revenue from this specific transaction, but may suffer a long-term decline in sales and revenues, because the consumer might become less loyal:
Perceived consumer manipulation in the digital world negatively affects loyalty towards the manipulator.
Taking the confirmation/disconfirmation paradigm (Oliver, 1980) into consideration, the manipulation of consumers can lead to negative disconfirmation, even though the consumer does not initially notice the manipulation (Román, 2010). In general, manipulation plays on decision biases or heuristics, increasing expectations about a product or service, while leaving performance unchanged. This might lead to negative disconfirmation and makes the consumer feeling dissatisfied (Román, 2010):
If consumer manipulation is perceived ex post (vs. ex ante), it leads to negative disconfirmation, thus leaving the consumer comparatively more dissatisfied.
Other factors, besides perception timing, can also have a negative impact (from a company’s perspective). As Lunardo and Mbengue (2013) suggest, the use of manipulative tactics leads to negative attitudes towards the applying organization and violates trust in the marketer (Greenspan, 2003; Sher, 2011). This can be explained by the equity theory (Adams, 1963). With regard to perceived consumer manipulation, if consumers perceive that they benefit less from a transaction than the supplier, they might feel treated unfairly, which has a negative effect on consumer satisfaction (Huppertz et al., 1978; Román, 2010). Furthermore, according to the attribution theory (Weiner, 1985), the manipulation of consumers, which restricts their autonomy, leads to consumers attributing negative outcomes to the supplier and thus negatively influence customer satisfaction and loyalty (André et al., 2018; Botti and McGill, 2011):
Perceived consumer manipulation has a negative effect on trust, consumer’s attitudes towards the organization, satisfaction and loyalty.
Taking the distinction between direct and indirect consumer manipulation into account, I further assume that if indirect manipulation occurs, the manipulation will be attributed to the third party carrying out the manipulation. Therefore, indirect manipulation might has fewer negative effects (for the organization) compared to direct consumer manipulation:
The negative impact for the organization is lower with indirect than with direct consumer manipulation.
However, in contrast to the previous argumentation, positive effects of perceived consumer manipulation might also occur. For example, if the manipulation satisfies second-order preferences (André et al., 2018; Jeffrey, 1974) or needs, of which the consumers were not aware of, consumers might be grateful and see the manipulation less critical. “Green Nudges,” for example, get consumers to make sustainable choices, thus contributing to close an intention-behavior gap:
Under certain circumstances, perceived consumer manipulation can lead to positive outcomes for consumers as well as societal welfare.
With regard to the different forms of manipulation (e.g., varying aggressive types of dark patterns), and by taking previous findings of Luguri and Strahilevitz (2021) into account, it can be assumed that the relationship between aggressiveness of consumer manipulation and its effectiveness might be modeled as an inverted U-shaped relationship. This proposition is based on the assumption that manipulation is most effective when it remains unnoticed. Thus, the effect of manipulation outweighs the effects of perceived consumer manipulation to a certain extent:
Manipulations are most effective when they remain unnoticed (usually when they are medium aggressive).
Furthermore, following the Fuzzy Trace Theory (Reyna and Brainerd, 1991), I assume that the age of a consumer is also relevant for susceptibility to manipulation. Fuzzy Trace Theory distinguishes between intuitive (gist-based) and analytic (verbatim-based) processing or decision making (Reyna and Brainerd, 1995, 2011). While gist-based processing captures and processes only the quintessence of information, verbatim-based processing captures the exact information in a quantitatively precise manner (Reyna, 2008). The older consumers are, the more experience they have in decision-making, therefore recognizing certain patterns more quickly. As a result, they tend to make intuitive (gist-based) rather than analytical (verbatim-based) decisions (Reyna and Brainerd, 1995, 2011; Reyna, 2008). Therefore, comparatively older consumers may be more sensitive to decision heuristics that can be exploited through manipulation. Proposition 8: Comparable older consumers are less likely to recognize manipulation as such and are therefore more susceptible to manipulation than younger consumers.
Likewise, the factor income, which according to Mani et al. (2013) is negatively correlated with cognitive capacity, could lead to consumers with comparatively lower incomes being more likely to use decision heuristics instead of cognitive reflection when confronted with a manipulation. Again, this may foster a higher susceptibility to manipulation:
Consumers with low incomes are more susceptible to manipulation compared to consumers with high incomes.
The ability to assess the trustworthiness and credibility of information encountered online (Guess and Munger, 2022) – known as digital literacy – also seems relevant. As outlined in the previous sections, consumer manipulation is characterized by, among other things, poor information quality or misleading information. A high level of digital literacy, i.e. the ability to assess the credibility of information in the digital world, should therefore lead to a lower susceptibility to manipulation:
Digitally literate consumers are more likely to recognize manipulation as such and are therefore less susceptible to manipulation.
Incorporating the derived propositions and the typology of dark patterns (Gray et al., 2018), further research questions and hypotheses emerge that open up exciting research opportunities. In this context, especially dark patterns, as a specific type of manipulation, provide a vivid example to illustrate the breadth of this research field. Thus, combining the five types of dark patterns with the different dependent variables on which consumer manipulation has an influence according to the propositions, already leads to broad set of hypotheses, that can be accessed by research (see Figure 3).
Accordingly, there is potential for future research in this area. For example, an interesting approach could be to investigate the extent to which the influence of several types of dark patterns on the dependent variables varies in different contexts. This could be fruitful e.g. for marketing management, to be able to identify possible differences between the various types of dark patterns in terms of their effects of action. To empirically validate the propositions of this paper, authors might use surveys or experimental designs. It would also be highly interesting to cooperate with, for example, an online retailer and analyze the impact of different dark patterns on consumer manipulation and related constructs.
Implications
Practical implications
For business practice, the conceptualization presented in this paper offers a first overview of possible negative consequences when using manipulative tactics to get consumers to behave in a certain way. Following the argumentation of this paper, negative consequences for organizations are the most severe when the manipulation is directly applied by the organization, is characterized by low changeability and is perceived after the purchase (ex post) through a third-party hint. However, based on the classification presented in this paper, one could assume that every manipulation leads to negative outcomes for organizations, and the outcomes only differ in their negativity. Furthermore, as for many organizations, their customers’ trust is a central asset, the overall recommendation would be to not use manipulative techniques at all. Considering ethical aspects, this recommendation is further strengthened, and organizations should treat their customers fairly, thus helping to retain consumer sovereignty. Hopefully, the introduction of the classification will lead marketers to rethink the way they interact with their (potential) customers.
Theoretical implications
The introduction and definition of the construct of consumer manipulation contribute to a better understanding of consumer decision-making in the digital context and will sensitize researchers and practitioners to the possible negative effects of consumer manipulation. This paper adds to the literature by introducing a sound definition of the term consumer manipulation and differentiating it from related constructs. This will allow researchers to analyze the phenomenon of consumer manipulation with more focus. Furthermore, the conceptualization of the construct of perceived consumer manipulation potentially helps to enable further investigation of the perception of dark patterns, as well as sensitivity to manipulation in an online context, as requested by Gray and colleagues (2021). Therefore, this paper opens new directions for future research on manipulative tactics.
Conclusion
This paper aimed to clarify the concept of consumer manipulation by developing a clear definition and classification of the consequences of consumer manipulation. Accordingly, consumer manipulation was defined as:
“[…] any action aiming to influence the decision-making process of a consumer in the digital world in a hidden, targeted, and often selfish manner with the understanding that the consumer has the feeling that he or she is free to make a decision, even though a large part of the decision was determined in advance by the organization’s activities.”
Furthermore, the developed classification indicates that the consequences of consumer manipulation might differ based on:
the source of manipulation;
its changeability;
the reasons for perception; and
perception timing.
To conclude, future research could empirically test the proposed classification as well as the propositions to validate the assumptions emerging from the classification and further deepen the knowledge of consumer manipulation.
