Understanding Authenticity in Digital Cause-Related Advertising: Does Cause Involvement Moderate Intention to Purchase?

The paper provides a survey understanding of two dimensions of perceived authenticity in digital cause-related marketing (CRM) display advertising and models the impact on consumers’ responses. It develops a model with a set of six hypotheses and tests them through a multivariate structural equation technique on quantitative data generated by a survey procedure on a UK-based consumer panel. The 465 online panel participants consisted of 60% males and 40% females between the ages of 18 and 35. The findings provide empirical evidence that the perceived donation amount and adcontext congruence are intertwined perspectives of authenticity in digital CRM display ads, and show that both of these dimensions provide input to conferring authenticity on the ads. However, the perceived donation amount accounts for a stronger effect than ad-context congruence. The paper finds firms’ altruistic motives an influential antecedent to the mediating role of attitudes towards the ad (AaD) as a fundraising tool. This is due to the nature of the impact of authenticity on the link with intention to purchase. Contrarily to expectation, the level of firms’ involvement with a social cause does not moderate intention to purchase from the consumers’ standpoint. The paper provides an interpretation of authentic perception in digital CRM advertising and proposes a composite model of the mechanism of this effect on consumers’ response. In DCRM advertising, increasing the perception of donation magnitude is a key driver to advertising success, and the level of association with a social cause is not a significant factor in segmenting and targeting consumers.

offers. To overcome this worthlessness, consumers are keen to find cues in advertising that provide real, honest and genuine offers to meet their expectations. Marketers increasingly use digital cause-related marketing (CRM) approaches that deliver digital ads with a promise to support a social cause contingent on the consumer performing a desired action, such as click-through purchases, 'liking' of a page, watching a video, sharing a post, taking a survey or even playing an online game. These cause-related advertising tactics have the potential to create brand awareness, promote a social cause and inspire consumers to demonstrate support for causes they care about (Varadarajan and Menon, 1988;Kotler et al., 2012;Engage for Good, 2018;Bergkvist and Zhou, 2019).
With the increasing convergence of price and quality on the internet, the deployment of cause-related advertising can attract enlightened consumers who are looking beyond consumption to participate in addressing social issues through their purchasing behaviour. However, research on the effectiveness of such advertising is limited and has mostly been described as a tool of online fundraising (Grobman, 2000;Harrison-Walker and Williamson, 2000;Husted and Whitehouse Jr, 2002;Hoefer, 2012). It is therefore essential to understand how consumers evaluate and respond to digital CRM ads, especially since digital advertising has become ubiquitous, intrusive and highly competitive. Prior research has indicated that authenticity is crucial to all advertising and marketing success (Beverland and Farrelly, 2010;Hallem et al., 2019;Becker et al., 2019), yet there is no common understanding of what authenticity entails (Grayson and Martinec, 2004). Hence, understanding the authenticity evaluation of digital CRM advertising can provide the basis for its appeal to digital consumers. The primary purpose of this paper was to determine what authenticity is in the context of digital CRM advertising. This paper addresses this purpose by establishing two objectives.
First, given that claims of authenticity are often schemed and designed by marketers with no collective meaning of what the term means (Arnould and Price, 2000;Grayson and Martinec, 2004), literature regarding CRM and display advertising was reviewed to identify salient dimensions of authenticity cues applicable to digital CRM advertising. The impact of perceived ad authenticity (PAA) on consumers' intention to purchase (IP) digital CRM advertising products was also empirically examined. Second, the paper aims to gain a deeper understanding of the PAA-IP link. For this a more detailed examination is necessary, and from a practical perspective, marketers need guidance on the intervening and influencing conditions under which the ad can have a positive impact on IP.
The following section evaluates the relevant literature and develops hypotheses to address these objectives.

Literature review and hypotheses development
The concept of advertising authenticity Authenticity has become one of the common words used in modern advertising, with both marketers and ad agencies convinced that execution of an authentic ad is key to effective advertising and brand success (Morhart et al., 2015;Hallem et al., 2019). Specifically, prior research has suggested that authenticity provides a competitive edge in crowded marketplaces (Hallem et al., 2019), stimulates brand trust (Anderberg and Morris, 2006), helps and moderates emotional attachment to a brand (Hallem et al., 2019), is key to the success of cause-related marketing partnerships (Kotler et al., 2012) and helps consumers find genuineness, truth and virtue within their mix of consumption goals (Michael and Beverland, 2010). Authentic advertising can help create personal brand resonance and helps overcome consumer scepticism about advertising. While researchers report these effects of authenticity, there is no clear definition of what constitutes authenticity in an advertising campaign, thus highlighting the multidimensional notion of authenticity (Leigh et al., 2006). For example, in advertising literature, authenticity has been linked to a spokesperson's trustworthiness (Stern, 1994). Some link it to a realistic plot (Deighton, Romer and MacQueen, 1989) and still others link it to an ad's accurate representation of the brand (Beverland, Lindgreen and Vink, 2008). In other words, there are several dimensions to advertising authenticity that are relatable to physical attributes or the brand essence as evaluated by consumers using their lenses of personal experience (Grayson and Martinec, 2004). Additionally, authenticity cues may be more prominent in one context and setting than another, and understanding what makes up authenticity in digital CRM ads should be key to understanding the effectiveness of the advertising tactic.

Perceived authenticity of digital cause-related ads
In the following paragraphs and sections, two dimensions and antecedents of authenticity are conceptualised in the context of digital CRM advertising, these being the perceived donation amount and the ad-context congruence, as depicted in Figure 1. These variables were evaluated based on an arduous literature review of these concepts within traditional CRM and display advertising literature. This approach is sound, given the potential impact of the combined effects of donation magnitude and ad congruence in digital CRM display advertising that could exhibit effects found in traditional CRM, which has only seen limited and dated survey investigation in the context of digital advertising (e.g. Grobman, 2000;Yue and Chaturvedi, 2000;Hoefer, 2012). Notwithstanding, digital CRM advertising is widespread, as depicted in Appendix 1, and should benefit from guidance in academic research.

Perceived donation amount in defining ad authenticity
Examples of digital CRM advertising, such as Laithwaite's offer to donate 3% of sales from cases of wine to the National Trust, and Arena Flowers' offer to donate 25% of the selling price of flowers to Cancer Research on their websites, indicate that donation amounts can vary widely, ranging from very small amounts up to 50% of the product price. In this transactionbased approach to CRM, the consumer implicitly triggers a donation to charity when making the purchase. The results of this paper indicate that donation magnitude is a primary cue of authenticity in digital CRM ads because it can demonstrate the degree of honesty, fairness and legitimacy to base an authenticity claim in the design of the advertising. In fact, the donation amount in digital CRM ads and traditional CRM promotions is a primary campaign design element because marketers are faced with the decision of how to counterbalance reduced profits with support of a social cause. Although the impact of the amount on a social cause in digital advertising has not been a subject of discussion in the literature, it is sound thinking to draw from CRM literature to suggest that a higher amount would generate a more favourable authenticity assessment, a 'warm glow' feeling (Andreoni, 1989) and a more positive feeling toward the advertisement (Pracejus and Olsen, 2004;Arora and Henderson, 2007;Chang, 2008;Muller et al., 2014). The essence of a non-profit organisation (NPO) collaboration with a brand promotion is contingent on the potential financial benefit to the social cause.
When consumers make purchases linked to supporting a charity or cause, they may first focus on themselves because the utility that they gain (e.g. moral satisfaction) determines how they respond to the advertising. The utility of shopping with a social cause in mind is determined by both the cost of the shopping and the moral satisfaction of contributing to a good cause (Koschate-Fischer et al., 2012). As such, an individual's favourable authenticity perception of a sales offer would create the feeling of moral satisfaction, or 'warm glow', which should increase with higher donation levels from the consumer's standpoint (Andreoni, 1989;Koschate-Fischer et al., 2012). By contrast, when the donation amount is small compared to the price of the product, consumers who are information adept may consider the contribution insufficient, inducing scepticism toward supporting an NPO in this way. Therefore, their interest in purchasing to benefit an NPO cause may be significantly weakened and thus translate to an insignificant purchase intention.
Since the donation amount is part of a digital CRM ad's executional or content cues, this implies that an authentic ad is one that is altruistic, genuine and fair with regard to the donation dimensions of the ad. In sum, the donation as an authenticity cue helps consumers identify the altruism, honesty and trustworthiness in digital CRM advertising.

Ad-website congruence in defining ad authenticity
The concept of congruence, or congruity, which means to 'fit', 'ink', or 'match-up' has been used to examine the relatedness between a brand and another entity. The underlying concept of congruence remains the same regardless of the terminological variants (Lee et al., 2017). In cause-related advertising, the congruence between a brand and social cause strengthens consumers' attribution of corporate altruism and credibility (Rifon et al., 2004), and product purchase intention (Becker-Olsen et al., 2006;Lafferty et al., 2007;Bigné-Alcañiz et al., 2012). Digital advertising research corroborates these findings by showing that thematic congruence between an ad and the site content renders a product more appealing (Flores et al., 2014), generates more click-through (Cho, 2003), increases customer awareness and generates better attention and recall compared to an incongruent ad (Rieger et al., 2015;Zanjani et al., 2011). An ad's context congruence also elicits favourable responses in terms of attitudes toward the ad (Choi and Rifon, 2002) and purchase intention (Jeong and King, 2010;Segev et al., 2014). Moreover, a much more recent investigation also revealed that congruence with the brand's essence has a positive effect on purchases (Becker et al., 2019). Taken together, these findings suggest that authenticity is a projection of consumers' beliefs, expectations and perspectives constructed from cues in the advertising, of which the perceived donation magnitude and ad-context congruence are critical elements in attributing authenticity and subsequent behaviours. This paper outlines a number of resulting hypotheses with the first stated as: H 1 : Perceived authenticity of the ad has a direct impact on consumers' intention to purchase.
Further to the above discussion on the role of donation magnitude and ad-context congruence of the ad in creating the perception of advertising authenticity, it is essential to understand the relative impact of the two dimensions. Intuitively, the underlying strategy of the ad is to elicit and support consumers' prosocial instincts when they are shopping. The non-profit partners of a firm have an expectation of a financial contribution more than anything else. The reputational risks for the social cause and the firm can be accounted for by the expected financial benefits from the advertising. Hence, this paper proposes that the magnitude of the donation offered to the social cause in designing the ad properties would constitute a more compelling feature than the perceived congruence of the ad context. Accordingly, the paper formally states that: H 2 : The impact of the perceived donation amount on the intention to purchase relative to the perceived ad-context congruence will be stronger.

The perceived altruistic attribution and attitudes towards the advertising
Consumers form beliefs concerning a brand's causal attribution or reasons for embedding social responsibility in advertising. Specifically, these attributions are related to the perceived sincerity or authenticity of the social responsibility actions (Hildebrand et al., 2011).
Social responsibility initiatives should be recognised as authentic to garner consumer support, and social responsibility programmes are perceived as authentic when they align with the brand's identity (McShane and Cunningham, 2012). Research has revealed that when there is dissonance between the brand's rhetoric and practice, it can substantially hurt the brand's image (Tsai, 2009) by creating negative perceptions about the brand (Hildebrand et al., 2011).
There are four components of altruistic motives within consumer responses to advertising: value-driven motives, strategic-driven motives, selfish motives and stakeholder-driven motives (Ellen, Webb and Mohr 2006;Groza et al., 2011;Lee et al., 2009). Although attributional altruistic motives have not been the subject of many studies in extended models as in the current paper, two theoretical frameworks have been applied to develop an understanding of the effects of attributed altruistic motives on consumer responses to digital CRM advertising. Firstly, by applying studies by Ajzen and Fishbein (1970) and Bagozzi (1981), which employed a commonly used attitudinal framework of cognition (beliefs) and donation (action) intention, the context suggests that all the components of the attributional altruistic motive, except for stakeholder-driven motivation, can positively impact attitudes and purchase intention. Secondly, the paper draws from from Bergkvist and Taylor's model of Leveraged Marketing Communications (LMC) that suggests a dual-path model of CRM persuasion effects. According to the model, CRM affects brand evaluations along an indirect transfer path, which is mediated by the attribution of motives and followed by a direct transfer path in which attitude towards the cause is transferred to the brand (Bergkvist and Zhou, 2018). The paper suggests that these frameworks can accurately clarify the effect of altruistic attributed motives and attitudes towards the ad in the context of the impact of perceived digital CRM ad authenticity. Hence, the paper proposes that: H 3 : Consumers' altruistic attribution mediates the effect of perceived ad authenticity on their intention to purchase. H 4 : Consumers' altruistic attribution influences attitudes towards the ad as a fundraising tool that affects their intention to purchase. H 5 : Consumers' attitudes towards the ad as a fundraising tool mediates the effect of perceived ad authenticity on their intention to purchase.

Influence of cause-involvement
The involvement factor has featured in several models investigating consumer behaviour in CRM (e.g., Hyllegard et al., 2011;Hajjat, 2013) and display advertising (e.g., Huang et al., 2010;Zanjani et al., 2011;Segev et al., 2014). These studies highlight that involvement plays a crucial role in moderating and explaining variable relationships. Hence, understanding how involvement might moderate the impact of perceived authenticity in CRM advertising will be essential to the development of more in-depth knowledge regarding how individual differences can impact the effectiveness of the ad. This paper evaluates involvement regarding how attention is paid to advertising with a social cause, which is indicative of a personal and enduring state of mind about a social cause. Involvement signifies personal values, needs, importance and interest, which modifies one's experiences and impacts one's involvement with the object (Zaichkowsky, 1986). Unlike other functional aspects of an advertisement, such as ad size and interactivity, the social cause associated with an ad can be deeply personal to a broad segment of the consumers. As such, the affective response toward an advertisement with an indication to support a social cause is likely to be amplified. Against this background, cause-involvement is expected to interplay with the perceived authenticity of the advertising to influence consumer responses. That is, the higher the relevance of the social cause, the stronger the scrutiny of the credibility and believability of the advertising. Also, a weak involvement with a social cause may not elicit an individual assessment of the brand's motive. Accordingly, the paper hypothesises that: H 6 : There will be a positive interaction of INVOL with PAA such that the greater the INVOL, the stronger will be the impact of the PAA on IP.

Study design, participants and procedure
Critical considerations in designing the study stimuli included the selection of different product categories that participants would typically buy online, low and high donation amounts offered by the ad and a charity website that would present the desired variability of the PAA variable. A low donation amount was set at 1%, and a high donation amount was set at 10%. For the products, plane tickets and toys were selected as products commonly purchased online. Hence, British Airways plane tickets and Argos 5-10-year-olds' toys were selected. The paper also selected Save the Children as the charity beneficiary of the CRM donation. The firms and the NPO are well known in this field of study, and they are regularly involved in CRM campaigns. Moreover, the use of reputable firms and an NPO was included to reduce or eliminate prior negative perceptions and de-legitimising advertising with CRM (Dean, 2003;Strahilevitz, 1999;Bigné-Alcañiz et al., 2012).
The study utilised a UK web-based consumer survey panel of 465 members, comprising 60% males and 40% females between the ages of 18 and 35. A standard market research agency provided the study participants, and participants were selected based on their experience with online shopping for plane tickets and toys from a variety of websites. Proprietary databases held by market research agencies are increasingly used in academic research because of their ability to provide high-quality responses quickly based on pre-determined selection criteria. Moreover, the web-based survey panel provided ecological validity of the study and had other advantages over traditional data collection methods, including low cost, time efficiency and avoidance of interviewer bias. The panel participants received an email invitation to the survey and were automatically entered into a points accumulation system that allows points earned in survey participation to be converted into shopping vouchers. In addition, participants could access the survey only once, and responses could be submitted only on completion of the survey. In effect, participants were presented with two samples  of Save the Children's website with a CRM banner ad advertising British Airways flight tickets, and a similar website site advertising Argos toys for children between the ages of 5 and 10 years old (Appendix 1). When the ads were presented to them, participants were asked several questions about the study constructs. The questions assessed the perception of a 1% and 10% donation offer of the price of the products, representing low and high donations in line with similar studies in traditional CRM context (Henderson, 2007;Chang, 2008;Muller et al., 2014). Questions also measured the attributed altruistic motive of the firm's attitude towards the CRM ad as a fundraising tool, intention to respond to the ad and consumers' involvement with the Save the Children cause. Using two different products, two different brands and two levels of donation in the form of a percentage of the price of the products ensures that there is variability in the perception of the ad-context congruence and donation amount that constitutes our proposed dimensions of ad authenticity. Moreover, the design would enable the results to be more generalisable in the broader context of digital CRM advertising.
A single variable that measured perceived ad authenticity (PAA) was computed by summing the ad-context congruence and perceived donation amount subscales, as conceptualised in this paper. Ad-context congruence was measured by adapting three items with bipolar adjectives from the scale proposed by Rifon et al. (2004). The perceived donation amount was measured on a three-item scale based on the approach described by Koschate-Fischer et al. (2012). Altruistic attributions were measured on a three-item bipolar scale following Ellen, Webb and Mohr (2006). Attitudes towards the ad were measured by adapting a four-point scale employed by MacKenzie et al. (1986), and the intention to purchase was measured by adapting a three-item bipolar scale (Yi, 1990). Finally, INVOL was measured on a three-item scale presented by Grau and Folse (2007). All the scaled items were measured on a seven-point semantic differential point scale. Appendix 2 presents the items on the scales used.

Assessment of the model measurement instruments' psychometric properties
To test the validity and reliability of each factor, our proposed model included a three-step data analysis approach: (1) data preparation; (2) exploratory factor analysis (EFA); (3) reliability test. Confirmatory factor analysis (CFA) was conducted using SPSS 26 and AMOS 26 statistical packages.
First, in preparing the data for analysis, it was necessary to excluded seven cases (<5%) from the data following data cleaning rules, as they exhibited evidence of unengaged responses. Since the factors are ordinal with seven intervals, extreme value outliers were not present. In addition, the factors were based on Likert-type scales, and there was no reason to exclude variables based on skewness (Gaskin, 2019). Thus, rather than testing skewness, the focus was on kurtosis (which was within +/-1.00). Linearity was tested by performing curve estimation regression with SPSS for all direct effects in the proposed model, which showed that the relationships between variables were sufficiently linear (i.e., all p-values were less than 0.05). In addition, the results of the homoscedasticity test (scatter plot of zPred on zResid) indicated that the proposed mediators were sufficiently homoscedastic. The variable inflation factors (VIFs) test for all the exogenous variables showed they all had a value of less than 2.0, meaning they were distinct from each other (Gaskin, 2019).
Second, to estimate the model, a factor analysis (EFA) was conducted using principal components with Promax rotation and factor loading of ≥0.30 to test whether the observed variables loaded together as expected, were adequately correlated and met reliability and validity criteria ( Table 1).
The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy and Bartlett's test, p = 000 were employed, and the communalities for each variable were sufficiently high (>.6), thus indicating that the study variables were adequately correlated for factor analysis (Gaskin, 2019). All items of the proposed model loaded satisfactorily except for item PDA3, which was removed to achieve a clean EFA, as depicted in the pattern matrix (Table 1). Additionally, the reproduced matrix had only 1% non-redundant residuals higher than 0.05, which further confirmed the adequacy of the variables and the sixfactor model. Importantly, all scales showed high internal consistency in the corresponding reflective indicators.
Third, analysis confirmed the convergent and discriminant validity of the data where for all factors the average variance extracted (AVE) was above 0.50 (Anderson and Gerbing, 1988;Hu and Bentler, 1999;Malhotra and Dash, 2011), and all factors demonstrated adequate discriminant validity because the diagonal values in Table 2 were greater than the correlations (Gaskin, 2019). The composite reliability (CR) of each factor attained the threshold of 0.70 (Nunnally and Bernstein, 1994;Hair et al., 2006). Hence it can be concluded that there was enough reliability for the factors to proceed with further analysis. Also, since all the latent factors showed high internal consistency of the corresponding reflexive observed items, composites were created for use in further analysis by averaging the values of the items.
The analysis confirmed the convergent and discriminant validity of the data where for all factors the average variance extracted (AVE) was above 0.50, and all factors demonstrated adequate discriminant validity because the diagonal values in Table 3 were greater than the correlations (Gaskin, 2019). The composite reliability (CR) of each factor attained the threshold of 0.70, hence it can be concluded that there was sufficient reliability for the factors to proceed with further analysis. Also, since all the latent factors showed high internal consistency of the corresponding reflexive observed items, composites were created for use in further analysis by averaging the values of the items.

Hypotheses testing results
The paper tested our mediation and moderation hypotheses using options 6 and 15, respectively, of Hayes' (2013) process macro in SPSS version 26 with a 5000-bootstrap resampling at the 95% confidence interval. The results of the significant model paths are depicted in Figure 2 and the full results are as detailed in Table 3. The paper tested the hypotheses while controlling for gender to remove potential confounding effects, since the literature suggests that gender is an influential variable in prosocial activities. Moreover, females have been found in most cases to show more favourable attitudes than men (Moosmayer and Fuljahn, 2010). Table 3 and Figure 2 summarise the results of the hypotheses testing. Interpretation of the direct, total and mediation effects of PAA -IP follow the regression, mediation and moderation bootstrapping process approach employed by Hayes (2013). The proposed final best-fit consumer response model was determined by model estimations in AMOS 26.
The aim of H 1 was to confirm the positive relationship between the two predictors of perceived ad authenticity (PAA) on the outcome variable (IP). The relationship was explored by performing regression analysis, and the results presented in Table 3 indicate that the regression coefficient was positive, and the relationship was statistically significant (.585(.052); p < 0.0005). Hence, H 1 was supported by the sample data. H 2 predicted the relative contribution of perceived donation amount (PDA) and ad-context congruence (CONGR) in PAA determining IP. The results confirmed the predicted stronger effect of PDA over CONGR (PDA: B = .444(.033); CONGR: B = .054(.042). Hence, H 2 was supported.
The goal of the mediation hypotheses was to investigate the mechanism through which the changes in PAA in digital CRM ads result in changes in IP. As indicated in Table 3 and illustrated in Figure 2, the significant mediatory paths confirm mediation (Hayes, 2013). In addition, attributed altruistic motives (MOTV) were a significant predictor of attitudes towards the ad (AAT): 359(.043) p < 0005. These results support the serial mediational hypotheses (H 3 , H 4 and H 5 ) of MOTV and ATT. However, since PAA was a significant predictor of IP before and after the inclusion of the mediators, the results were indicative of partial mediatory effects for both MOTV and ATT on PAA-IP (Hayes, 2013;Gaskin, 2019). Hypothesis H 6 predicted how the consumers' involvement (INVOL) with the social cause might strengthen or weaken the PAA-IP relationship. The results show that INVOL does not significantly vary across situations in digital CRM advertising, as predicted (B: -.044(.043), P = .253). Instead, INVOL seems to dampen the positive PAA-IP link, as illustrated in Figure 3. This illustrates that cause-Involvement dampens the positive relationship between perceived ad authenticity and intention to purchase (IP).
To propose an effective consumer response model, the paper attempted to fit the original conceptual structural model of the variables to the data. Unfortunately, due to the presence of serial mediators, the model would have had zero degrees of freedom and thus fit perfectly.
However, the paper presents this model because it represents what happens (as already proven) when MOTV and ATT both have direct effects on IP. The model would be perfect (as has been shown) when MOTV and ATT work in the same way towards IP. To achieve some degree of freedom (Goldsmith et al., 2000;Bagozzi and Yi, 2012), the paper considered the best alternate model to fit the sample data (Goldsmith et al., 2000;Schreiber, 2008). An examination of the model parameter estimates (Figure 2) and fit statistics indicated that the path from MOTV to IP was relatively weak, since MOTV was found to have less effect on IP than ATT (Goldsmith et al., 2000). Hence, this path was eliminated.
This modification is consistent with the attitude-toward-the-ad literature that specifies the Dual Mediation Hypothesis (MacKenzie et al., 1986), which shows only an indirect effect of ' attitudes' on behavioural outcomes. Furthermore, according to the Elaboration Likelihood Model (ELM) presented by Petty et al. (1983), the peripheral route to persuasion is the path where the digital CRM offer could be most compelling, and the effects are most closely associated with attitudes towards the ad. The goodness-of-fit statistics for the revised model (Figure 4), as recommended by several authors (Anderson and Gerbing, 1988;Hu and Bentler, 1999;Schreiber), indicated a reasonably good overall fit (Goodness-of-fit indexes: CMIN/DF = 2.260; p = 000; GFI = 946; CFI = .886; AGFI = .922; RMR = .204; NFI = .816; RMSEA = .052; PCLOSE = .349).

Summary and conclusions
The goal of this paper was to identify and examine the impact of two dimensions of authenticity of digital CRM ads on consumer responses while also proposing a composite theoretical model for advertising effectiveness. There are several reasons for examining the impact of the authenticity of digital CRM ads with cause-involvement as a moderator. First, advertising authenticity is considered to be a significant driver of advertising. Yet, there is no common understanding of what ad authenticity involves (Morhart et al., 2015;Becker et al., 2019). Second, because authenticity can help consumers find how to attain personal goals (Michael and Beverland, 2010), it is crucial to understand how digital CRM advertising resonates with consumers and influences consumption behaviour. Third, and following on from the preceding point, it is argued that different personal goals and standards enable people to find authenticity in a range of objects and consumption situations that others may deem as fake (Rose and Wood 2005;Michael and Beverland, 2010). Therefore, it is crucial to understand how consumers view and respond to digital CRM advertising.
By identifying donation magnitude as a critical element in ad design with respect to authenticity, this paper show that consumers' assessments of genuineness, reality and truth on how to attain the social goal of helping others is based on that factor. This result also recognises previous work within conventional CRM research which suggests that the donation magnitude has a significant effect on the brand evaluation (Koschate-Fischer et al., 2012;Chang, 2012;Folse et al., 2014). In addition, it determines the credibility, honesty, trustworthiness and transparency of the advertising (Webb and Mohr, 1998;Moosmayer and Fuljahn, 2013). In addition, in line with the congruity theories of contextual priming (Yi, 1990(Yi, , 1993 and cognitive interference (Furnham et al., 2002), this paper proposes that the ad-context congruence as a component of perceived ad authenticity would reflect findings in various contexts examined in CRM research, where 'relevance', ' congruity' and 'fit' (as they have been variously termed) provide cues that consumers use to evaluate company trustworthiness when linking advertising to a social cause (Rifon et al., 2004;Becker-Olsen et al., 2006;Ellen, Webb and Mohr 2006;Lafferty et al., 2007;Bigné-Alcañiz et al., 2012). Besides, when considering congruence as an authenticity element in digital CRM advertising, this paper recognises previous works that suggest that ad-context congruence can increase attention, memory and retention (Choi and Rifon, 2002). Hence, the findings of this paper validate these works in the context of digital CRM display advertising and show that donation magnitude and ad-context congruence define authenticity indicators of the ad. By evaluating the relative impact of the donation amount and ad-context congruence, this paper also contributes to the literature by showing that donation magnitude has a stronger impact on consumers' authenticity evaluation of the ad than ad-context congruence. A valuable insight obtained from the proposed model is that PAA has an impact on consumer behaviour through altruistic attribution of the firm (MOTV) in the advertising and attitude towards the ad as a fundraiser (AAT), because attitudes mediate the positive PAA-IP link (Figure 4).
The proposed model did not show support for influence involvement (INVOL), which is the intangible attitude to values that are derived from the relevance and importance that people hold towards the social cause. The effectiveness of ads has been found in various context in CRM research (Hyllegard et al., 2011;Koschate-Fischer et al., 2012;Hajjat, 2013) and display advertising (Huang et al., 2010;Zanjani et al., 2011;Segev et al., 2014). For years advertising scholars have argued that involvement plays a significant role in moderating and interpreting variable relationships (Muehling et al., 1993;Belch and Belch 2012), affecting the level of arousal and preparedness to approach a display ad on a website.
The counterintuitive results thus provide fresh insights into the role of involvement and the nature of digital CRM display advertising that operationalises concepts of conventional CRM within the digital space. More specifically, and in contrast to the wellestablished ELM (Petty et al., 1983), the findings suggest that participants who are more involved with the study of a social cause (e.g., Save the Children) do not process the prosocial message of the ad strongly enough to stimulate more favourable cognitive elaboration of the ad as expected. Involvement was found to marginally dampen the positive PAA-IP relationship.

Managerial implications
It is widely claimed that authenticity is a major driver of advertising success (Hallem et al., 2019;Becker et al., 2019). However, it is not clear what authenticity means in different contexts and situations. This research presents an initial attempt to shed light on the concept of authenticity in the context of digital CRM advertising and its impact on consumers' intention to purchase. The paper provides evidence that the cumulative effect of perceived donation amount and ad-context congruence can define the authenticity of the ad. The identification of the two dimensions implies that marketers can use the term ' authenticity' more specifically in internal and external communications when, for example, working with a digital creative designer. Our research thus helps create the basis for more fruitful and targeted communications.
Second, advertisers must appreciate consumers' altruistic attribution motives in the ad. The results of the paper show that consumers attempt to make attributions of the firm's support to the social cause. Thus, any managerial decision that improves perceptions of company trustworthiness (e.g., increased donations) will also increase perceptions of the altruistic image as more legitimate, genuine and sound, which significantly improves consumer attitudes and responses towards the ad. Even the ad congruence with the NPO's cause website context, renders the ad more legitimate and genuine. This contextual coherence cue makes it possible for consumers to form favourable attitudes towards the ad, leading to the intention to support the advertising.
Finally, regarding the interplay of authenticity and consumers' personal attributes, the results of this paper did not support claims that consumers with greater involvement in social causes would demonstrate a higher intention to purchase given a higher degree of perceived authenticity. This finding can be more broadly taken to suggest that the ad could be used for effective targeting of a more diverse consumer segment in terms of differences in connection with the social cause.

Limitations and future directions
While the present paper extends the understanding of existing advertising authenticity and CRM research in several ways and provides essential theoretical and managerial implications for digital advertising, a significant limitation should be mentioned. Recognising this should help refine future research efforts.
A significant limitation to this paper is that the design and placement of the ad stimuli on a charity website rather than the commercial brand website renders the study more focused on the charitable cause than the commercial brand. As a result, the proposed model may not be generalisable to a context where the brand is a central focus.
There are two possible directions in which this study can be extended. First, given that this paper found the effect of authenticity is only partially mediated by attributed firm altruistic motive and attitudes towards the ad as a fundraiser there could be other significant mediators of authenticity to be identified, which considered together can fully explain the positive relationship between authenticity and intention to purchase.
Second, future research should attempt to replicate these results and determine the differential impact of donation amount and congruence in models containing a variety of additional moderating constructs, such as 'familiarity' with the brand or social causes.
Despite the limitation indicated, the results in this paper demonstrate that the perceived donation amount and the relevance of the ad constitute the authenticity of digital CRM ads, and the altruistic motives attributed to the brand and attitudes towards the advertisement as a fundraiser determine the intention to purchase to a large extent.