Impact of movie-watching on cross-selling revenue in shopping malls: Implications for post-pandemic recovery

Abstract

The COVID-19 pandemic has prompted consumers to postpone in-store purchases and shift to online shopping, decreasing overall retail spending. As the global lockdown eases and shoppers return to offline stores and malls, shopping malls seek to facilitate in-store consumer experiences at entertainment facilities. Given the importance of entertainment facilities to the cross-selling revenue of shopping malls, we investigate the relationship between consumers' movie-watching and subsequent (hedonistic) grocery purchases. Our analysis of a panel database that includes transaction details at the consumer level shows that movie-watching is negatively correlated with consumer spending. Furthermore, consumers’ psychological states induced by content and contextual factors moderate the negative effect of movie-watching; after watching a fantasy (as compared to thriller) genre or watching a movie in a less crowded theater, consumers are more likely to make hedonic purchases at the supermarket. These findings provide theoretical and managerial implications for retail operators exploring cross-selling opportunities post-pandemic.

1. Introduction

The COVID-19 pandemic has profoundly impacted the retail industry, as the risk of disease transmission inside retail stores has changed consumers’ shopping behavior (Delasay et al., 2022; Donthu and Gustafsson, 2020; Laato et al., 2020; Pantano et al., 2020). As a result, consumers have postponed in-store purchases in favor of online shopping (Lynch, 2022; Renner et al., 2020). In the United States, 17% of consumers avoided visiting their favorite stores in person shortly after the pandemic began, and in China, grocery transactions fell by 30% (Briedis et al., 2020). Furthermore, recognizing the economic risks of the pandemic, consumers have opted to save for financial stability. Likewise, an Australian consumer survey has found that 55% of the respondents saved their household income and refrained from spending (Wegner et al., 2021). Such changes in consumer purchase behavior during the pandemic call for the post-pandemic sustainability of many existing face-to-face retail models that rely heavily on in-store experiences, amid a remarkable recovery since the imposed lockdowns have eased (Briggs, 2021; Eger et al., 2021). This study aims to draw implications for the post-pandemic era by examining how shopping malls can increase profits by leveraging cross-selling opportunities with movie theaters.

Shopping malls play an important role in the retail landscape by bringing together a wide range of consumer segments and retailers. To retain consumers, shopping malls focus on offering spaces for entertainment and recreational activities, such as exhibitions, aesthetic displays, pop-up events, and movie theaters, in addition to retail stores. Conventional wisdom states that such spaces increase consumers’ excitement and enjoyment levels (Farrag et al., 2010; Wakefield and Blodgett, 1999). By introducing entertainment facilities and eliciting consumer intimacy, shopping malls increase foot traffic and generate more revenue from retail stores (Kamakura, 2008). As the spread of COVID-19 has limited the number of consumer visits, shopping malls have been paying more attention to crowd-drawing spaces to extract more cross-selling retail revenue per consumer visit (Standish et al., 2021; Verdon, 2022). However, little is known about the extent to which consumer entertainment experiences affect their subsequent purchasing behavior at retail stores in shopping malls.

While the literature on cross-selling is extensive and well-established, cross-selling in shopping malls constitutes a research gap. Extant literature has focused on the antecedents and consequences of cross-selling, such as favorable perceptions (Dahana et al., 2020; Evanschitzky et al., 2017), behavioral loyalty (Béal et al., 2019; Reinartz et al., 2008), and the role of emotions and moods associated with overall consumption in shopping malls (Haq and Abbasi, 2016; Helmefalk and Hultén, 2017). These studies, however, did not examine the relationship between (non-retail) experiences and retail revenue in shopping malls. This study seeks to address the specific gap in the literature on cross-selling in shopping malls by investigating the following research objectives. First, we examine consumers' purchase behavior across entertainment facilities (i.e., movie theaters) and retail stores (i.e., supermarkets) in shopping malls. Second, we assess the extent to which consumers’ psychological state, arising from the content and contextual factors of movie-watching, affects their spending on hedonic product categories at retail stores.

To achieve our research objectives, we first model consumer spending across movie theaters and supermarkets in shopping malls, revealing the potential cross-selling revenue effect of offering entertainment facilities. We focus on movie-watching in that movie theaters are an important channel for large retailers to attract consumers (Yohn, 2020), and the market is rapidly recovering from the COVID-19 crisis (PwC, 2022). To further assess the effect of movie watching on retail revenue, we build a system of two simultaneous equations: (1) the first captures the consumers’ decision to watch a movie in the mall, and (2) the second characterizes their spending on hedonic product categories in retail stores. We hypothesized that when consumers watch a movie, they experience feelings or states of arousal that are likely to induce hedonic purchases. The proposed models are calibrated to a rich proprietary panel database obtained from a leading shopping mall in the Middle East, which includes transaction details at the consumer level over time.

This study makes several contributions to the literature. Contrary to the conventional wisdom that entertainment facilities help the overall retail revenue of a mall (Kenyon, 1997; Kim, 2001), our findings indicate that entertainment experiences can negatively influence such revenue gains. At the same time, we suggest a “silver lining” of offering entertainment experiences such that the psychological states induced by both content and contextual factors associated with the experiences could increase consumer spending on hedonic products at retail stores. As such, this study enriches the current limited understanding of cross-selling in shopping malls and provides managerial insights into generating substantial revenue in offline retailing in the post-pandemic era.

The remainder of this paper is organized as follows. Section 2 outlines the conceptual framework and proposed hypotheses. We also present a pertinent review of the relevant literature in this section. Section 3 provides a description of the methodology, including the data employed, model specifications, and empirical results. Finally, Sections 4, 5 present the findings of the study, its limitations, and directions for future research.

2. Conceptual framework and hypothesis development

In diverse retail settings, cross-selling refers to the total number of product categories that a firm sells to consumers (Li et al., 2011; Venkatesan and Kumar, 2004). Research has demonstrated that retail operators benefit from cross-selling to increase sales volume per consumer. Specifically, offering entertainment spaces in shopping malls encourages consumers to stay longer and visit multiple retailers, thereby increasing cross-selling revenue (Kamakura, 2008). This is because entertainment facilities in which the service environment plays an important role, such as movie theaters and hotels, generally induce emotional arousal and pleasure and improve consumer evaluations of shopping experiences (Mano and Oliver, 1993; Pham and Sun, 2020). This study thus examines the role of entertainment experiences in shopping malls and proposes a comprehensive framework that profiles mall visitors based on their behavioral characteristics (see Fig. 1 ). In this framework, we divide mall visitors into two groups: those who visited a movie theater and those who did not. The former group is further classified into two groups: those who made purchases at retail stores before or after watching a movie and those who did not. This allows us to identify the extent to which movie-watching contributes to retail revenue as well as the mechanism underlying it.

Fig. 1.

Classification of consumers in stages according to their behavior.

2.1. Movie theaters as a tool for inducing cross-selling in malls

As shopping becomes increasingly recognized as a leisure activity, the existence of entertainment facilities such as movie theaters increases the inflow of consumers into shopping malls (Jansen-Verbeke, 1987; Kesari and Atulkar, 2016; Martin and Mason, 1987). The rationale is that consumers who visit a mall to watch a movie recognize their unmet utilitarian or hedonic needs and choose the best alternatives to satisfy these needs by visiting stores (Levy et al., 2012). In turn, shopping malls increase overall revenue through cross-selling opportunities. However, it is expected that retail revenue gains induced by consumer movie experiences may not be as favorable for planned visits for which the sole purpose is to watch a movie. Given that planned shopping behavior is performed to address previously recognized problems or with purchase intentions formed before entering a store (Meng and Xu, 2012; Piron, 1993), consumers having a specific objective to visit the mall will be less likely to benefit from cross-selling. Based on this discussion, we predict that consumers watching a movie spend less at other retail stores in shopping malls and formulate the following hypothesis.

H1

A consumer who watches a movie will spend less during subsequent retail transactions in the mall than one who does not watch a movie.

2.2. Movie-watching and its affective consequences on grocery shopping

During a shopping mall visit, consumers are assumed to spend time sequentially and make a series of decisions in which one choice follows another; for example, they may watch movies, shop, eat food, and purchase groceries. Due to its sequential nature, prior decisions or choice outcomes will influence subsequent preferences in a consumption sequence (Dhar and Simonson, 1999). The initial choice leads to the activation of specific goals, which guide subsequent choices consistent with the same focal goal unless the goal progresses to activate an alternative goal. Once a specific goal is activated or constructed, it influences the choice of subsequent actions to help achieve the goal (Markman and Brendl, 2000). When the initial choice signals a commitment to a focal goal (e.g., a healthy entrée), people tend to make consistent subsequent choices (e.g., skipping a dessert) (Fishbach and Dhar, 2005).

Given that a movie offers hedonic experiences such as enjoyment, fun, and excitement (Hirschman and Holbrook, 1982; Okada, 2005), consumers are likely to activate a hedonic goal upon watching a movie. To fulfill this goal, they are expected to engage in hedonic purchases that maximize pleasure when consumed (Dugan et al., 2021) during subsequent retail transactions. When shopping for grocery items after watching a movie, for example, consumers may spend more on items containing a large amount of fat or sugar (Drewnowski et al., 1989; Oliveira et al., 2018). We thus predict that watching a movie positively impacts subsequent hedonic (grocery) consumption by changing consumers’ psychological states. Therefore, we formulate the following hypothesis.

H2

A consumer who has watched a movie spends more on hedonic (grocery) products at the mall compared to a consumer who has not yet watched a movie.

Literature has also suggested that a positive psychological state expands and builds a possible repertoire of future behavior (Fredrickson, 1998). As such, changes in psychological state after watching a movie are expected to influence subsequent consumer behavior in shopping malls (Sherman et al., 1997; Das and Varshneya, 2017). In evaluating the effect of movie-watching on subsequent hedonic retail purchases among moviegoers, therefore, a movie genre is one of the critical elements to reflect consumers' emotions in that it prompts different feelings and has different consequences for shopping behavior. Fantasy movies, for example, could evoke playful attitudes and create a climate of pleasure, escape, and relaxation, similar to consumption experiences in fantasy settings such as Las Vegas (Belk, 2003) and Disney amusement parks (Fjellman, 1992). In literature, fantasy is defined as a consumer's “mental activities of building an illusory world or reality that meets an individual's psychological needs and interests” (Wu and Holsapple, 2014) and is often seen as the creation of new worlds, promoting freedom from the limits of what is known and believed (Kozinets et al., 2004). As such, it has been connected to departure from the real world, enabling relaxation and renewal (Belk and Costa, 1998). In contrast, a thriller induces an affective state associated with uncertainty regarding an emotionally significant event that motivates future-oriented prediction and the desire for resolution (Lehne and Koelsch, 2015), creating strong psychological tension (Schmälzle and Grall, 2020).

Further, emotions are associated with consumers' mindsets. In particular, positive emotions such as happiness are associated with a promotion-focused mindset (Förster et al., 1998), and hedonic products align with this focus (Chitturi et al., 2008). In our context, watching “happy” movies, such as fantasy films, evokes positive feelings, leading to promotion-focused mindsets. As such, we anticipated that watching a fantasy movie prompts a positive emotional state that satisfies viewers’ fantasies and leads consumers to spend more on hedonic products. On the other hand, anxiety is associated with a strong prevention-focused mindset, and utilitarian products align well with this mindset. In this context, watching “tense” movies, such as thrillers, elicits negative feelings, leading to a prevention-focused mindset. Thus, consumers who watch thrillers are expected to spend less on hedonic products. Therefore, we hypothesize the following.

H3

The effect of movie-watching on moviegoers' spending on hedonic (grocery) products at the mall is moderated by the genre of the movie.

In addition to the movie genre, the physical space inside the movie theater, i.e., space configuration, could affect consumers' movie-watching experiences and emotional states, affecting their subsequent behavior. Social impact theory (Latané, 1981) suggests that people are impacted by the real, implied, or imagined presence or action of another person or group of people. As such, social density, i.e., the number of individuals in a given area (Baumeister and Vohs, 2007), has been considered an important factor in consumer retail purchases in that it reduces shopper satisfaction, desire to leave the store earlier (Eroglu et al., 2005; Hui and Bateson, 1991), and consumer perceptions of uniqueness (Xu et al., 2012). Likewise, a crowded environment that restricts one's personal space decreases participants' feelings of comfort, producing stress (Collette and Webb, 1976) and a lack of control (Hui and Bateson, 1991); thus, consumers seek to minimize crowding (Argo et al., 2005; Maeng et al., 2013). High social density also induces defensive responses (Felipe and Sommer, 1966; Sommer and Becker, 1969) due to personal space violations. For example, Oldham and Fried (1987) showed that office workers exhibit more withdrawal behavior in an office with overcrowded workstations. Being in a socially crowded situation leads to physiological symptoms of anxiety, high arousal, and low experience of pleasure (Schaeffer and Patterson, 1980).

In our context, having fewer empty seats in a movie theater could create a socially crowded environment, decrease consumers' feelings of comfort, and elicit negative effects such as stress. As we stated in H₃, anxiety and avoidance motivations are associated with a strong prevention-focused mindset (Förster et al., 1998), and utilitarian products align well with this focus on prevention (Chitturi et al., 2008). On the other hand, happiness and approach motivation is associated with a promotion-focused mindset, and hedonic products align with this focus. In turn, we predict that fewer empty seats in movie theaters activate the avoidance system, resulting in a prevention-focused mindset. In contrast, a less socially crowded environment accompanied by more empty seats could induce consumers’ feelings of comfort and trigger the approach system, resulting in a promotion-focused mindset. Therefore, we propose the following hypothesis.

H4

A moviegoer who watches a movie in a less crowded environment will spend more on hedonic (grocery) products at the mall than one who watches a movie in a more crowded environment.

3. Methodology

3.1. Data description

This study uses transaction details for each consumer from January 2017 to June 2018, obtained from shopping malls operated by a large conglomerate in the Middle East. The data consist of 2,804,290 supermarkets and 2,094,894 movie transactions made by 59,509 consumers during this period. Supermarket transaction data include 38,899,980 stock-keeping units (SKU) across 52 major categories; the movie transaction data include 5,854,125 ticket purchases across 1076 movies.1 These two datasets can be matched based on the consumer ID and shopping mall visit dates, allowing us to identify both movie and retail transactions during a single visit to the mall. The operationalization descriptive statistics of the focal variables are presented in Table 1, Table 2 , respectively.

Table 1.

Variable operationalization.

Variable	Notation	Variable type	Operational definition (Measurement)
A. Movie variables
Movie	$Movi{e}_i$	Binary	Whether a consumer i who visited the shopping mall watched a movie at a movie theater adjacent to or inside the mall on the same day
Content variables
MovieGenre	$Movi{eGenre}_i$	Categorical	The genre of movie that consumer i watched
MajorStudio	$MajorStudi{o}_i$	Binary	Whether the movie the consumer i watched was produced by one of the global filmmakers (i.e., Universal Pictures, Warner Bros, Columbia Pictures, Walt Disney Pictures, Marvel Studios, Paramount Pictures, and Sony Pictures)
Context variables
Quantity	$Quantit{y}_i$	Numerical	No. of tickets booked by the consumer i at the same time
SeatsAvailable	$SeatsAvailabl{e}_i$	Numerical	No. of empty seats in a movie theater when the consumer i is watching a movie
B. Store visited
LeisureEnt	$LeisureEn{t}_i$	Binary	Whether the consumer i who visited the shopping mall used internal or adjacent leisure and entertainment facilities other than the movie theater
Fashion	$Fashio{n}_i$	Binary	Whether the consumer i who visited the shopping mall used an internal or adjacent fashion store
C. Segment
HotelVisit	$HotelVisi{t}_i$	Binary	Whether the consumer i who visited the shopping mall is a hotel consumer
Ecoupon	$Ecoupo{n}_i$	Binary	Whether the consumer i who visited the mall used an e-coupon
Wi-Fi	$Wif{i}_i$	Binary	Whether the consumer i who visited the shopping mall used the store Wi-Fi on the same day
D. Demographic
Gender	$Gende{r}_i$	Binary	Binary variable indicating consumer i's gender (Female = 0, Male = 1)
Age	$Ag{e}_i$	Numerical	Consumer i's age
Dependents	$Dependent{s}_i$	Numerical	No. of dependents on consumer i
Monthly salary	$Salar{y}_i$	Numerical	Consumer i's salary level (Unit: DH)
Employment status	$Employmen{t}_i$	Categorical	Category variable representing the employment status (i.e., Salaried, Self-employed)
Occupational Status	$Occupatio{n}_i$	Categorical	Category variable representing the occupational status (i.e., Officer, Manager, Higher than manager)

Table 2.

Summary of descriptive statistics.

Variables	Observations	Mean	Std. dev.	Min	Max
Movie	2,094,894	0.229	0.420	0	1
MajorStudio	2,094,894	0.102	0.302	0	1
Quantity	2,094,894	2.790	1.582	1	22
SeatsAvailable	2,094,894	67.143	85.483	0	535
LeisureEnt	2,804,290	0.117	0.322	0	1
Fashion	2,804,290	0.036	0.187	0	1
HotelVisit	2,804,290	0.010	0.098	0	1
Ecoupon	2,804,290	0.219	0.414	0	1
Wi-Fi	2,804,290	0.380	0.485	0	1
Gender	59,509	0.660	0.474	0	1
Age	59,509	39.593	9.291	18	71
Dependents	59,509	2.185	1.400	1	9

3.2. Empirical analyses

To obtain initial evidence of the impact of movie-watching on mall revenue, we first examine whether consumers watching a movie generally spend more at retail stores, using a Tobit regression:

$$Y=\max \left\{Y^{*},0\right\}\mathrm{w}\mathrm{h}\mathrm{e}\mathrm{r}\mathrm{e}{\mathrm{Y}}^{*}=X^{\prime }\beta +e\text{.}$$ (1)

where $Y^{*}$ refers to the latent variable associated with dependent variables: the log-transformed purchase amounts and retail (grocery) purchase amounts in the shopping mall. X includes transaction details, such as whether to book a movie at least seven days before the shopping mall visit and the number of tickets purchased. We also control for consumer-specific characteristics, such as demographics, employment, and occupational status, as well as behavioral segments determined by the mall. $e\sim N\left(0,{\sigma }^2\right)$ is the error term.

Our main model is the proposed two-stage model which aims to examine the extent to which consumer (ex-post) spending on hedonic purchases is induced by their movie-watching experience. The first stage employs a Probit regression to examine consumers’ decisions to watch a movie. For consumer $i$ with occupational status $j$,

$$P\left(Movi{e}_i=1|X\right)=\frac{\exp \left(u_i\right)}{1+\exp \left(u_i\right)}$$ (2)

where $Movi{e}_i=1$ if $u_i>0$, and $Movi{e}_i=0$ if $u_i\le 0$ for consumer i.

$$u_i={\beta }_0+{\beta }_{1}Frequenc{y}_i+{\beta }_{2}Ag{e}_i+{\beta }_{3}Gende{r}_i+{\beta }_{4}Employmen{t}_i+\sum _{j=5}^6{\beta }_{j}Occupatio{n}_{ij}+{\beta }_{7}Salar{y}_i+{\beta }_{8}Dependent{s}_i+{\beta }_{9}Weeken{d}_i+{\epsilon }_i\text{,}$$

where $Ag{e}_i$, $Gende{r}_i$, $Salar{y}_i$, $Dependent{s}_i,Employmen{t}_i,andOccupatio{n}_{ij}$ refer to consumer-specific characteristics such as age, gender, monthly salary, number of dependents, employment status, and occupational status, respectively. $Frequenc{y}_i$ refers to the average number of movies consumer $i$ watches per year as an instrumental variable. $Weeken{d}_i$ indicates whether consumer $i$ visits the movie theater on the weekend. ${\epsilon }_i$ is the error term.

In the second stage, we investigate consumer spending on hedonic products in supermarkets, conditional on their experiences in the movie theater. We set the dependent variables as log transformations of the total amount of hedonic purchases and hedonic grocery purchases for each transaction in the shopping mall. The former includes product purchases with hedonic characteristics, such as jewelry and accessories, perfumes, hi-fi equipment, and groceries, such as chocolate, enriched bread, and cakes. The latter includes only products that have the characteristics of hedonic purchases within the grocery category, which generally includes products that contain high levels of sugar or fat (Drewnowski et al., 1989). The proposed model is as follows:

$$\ln \left(Hedoni{c}_i\right)|\left(Movi{e}_i=1\right)={\beta }_0+{\beta }_{1}Afte{r}_i+{\sum }_{j=1}^2{\beta }_{2,j}MovieGenr{e}_{i,j}+{\beta }_{3}MajorStudi{o}_i+{\beta }_{4}SeatsAvailabl{e}_i+{\beta }_{5}Quantit{y}_i+{\sum }_{j=1}^2{\beta }_{6,j}Afte{r}_i\times MovieGenr{e}_{i,j}+{\beta }_{7}Afte{r}_i\times SeatsAvailabl{e}_i+{\beta }_{8}LeisureEn{t}_i+{\beta }_{9}Fashio{n}_i+{\beta }_{10}HotelVisi{t}_i+{\beta }_{11}Ecoupo{n}_i+{\beta }_{12}Wif{i}_i+{\tau }_i+{\beta }_{IMR}{\lambda }_i+{\epsilon }_i\text{,}$$ (3)

where $Afte{r}_i$ refers to whether consumer $i$ shopped after watching the movie (1 if watched a movie before visiting retail stores; 0 otherwise) and $MovieGenr{e}_{i,j}$ refers to the genre of the movie (j refers to the index of the movie genre).2 $MajorStudi{o}_i$ indicates whether the movie was produced by a globally famous production studio. $SeatsAvailabl{e}_i$ refers to the number of empty seats available upon watching the movie, and $Quantit{y}_i$ refers to the number of tickets consumer $i$ purchased. The model also incorporates the retail stores consumer $i$ visited and other transactional attributes such as e-coupon or Wi-Fi usage during the mall visit.

To address potential endogeneity in movie-watching and the subsequent hedonic (grocery) purchase amount in the shopping mall, the proposed model adopts the Heckman two-stage model (Heckman, 1979) in which the Inverse Mills Ratio (IMR), ${\lambda }_i\text{,}$ adjusts the selection bias resulting from non-randomly selected samples (Thirumalai and Sinha, 2013). For identification purposes, we construct an instrument variable, i.e., the average number of movies consumers watched per year (Angrist and Pischke, 2009). For our instrumental variable to be valid, we ensure that the variable affects consumers’ movie-watching (Wald ${\chi }^2=311.21$; $p<0.01$), but not the hedonic (grocery) purchase amount in the shopping mall (with the raw correlation being essentially zero at 0.0059).

4. Results

Table 3 reports the results from the Tobit model estimating the relationship between watching a movie and the total (grocery) purchase amount at supermarkets in the mall (Equation (1)). The parameter estimates indicate that consumers who watch a movie, compared to those who do not, spend less at supermarkets ($\beta =-0.098;p<0.01)$ and grocery spending ($\beta =-0.171;p<0.01)$, supporting H₁. The results also show that consumers visiting hotels and buying fashion items spend more, whereas those using Wi-Fi services spend less.

Table 3.

Model parameter estimation results.

Variables	DV: ln(Purchase amounts)				DV: ln(Purchase amounts of groceries)
	Coef.	Std. err.	Coef.	Std. err.	Coef.	Std. err.	Coef.	Std. err.
Movie	−0.098***	0.008			−0.171***	0.008
Quantity			0.048***	0.006			0.035***	0.006
StoreVisited
LeisureEnt	0.227***	0.002	0.199***	0.019	0.229***	0.002	0.188***	0.020
Fashion	0.193***	0.005	0.337***	0.046	0.238***	0.005	0.330***	0.049
Segment
HotelVisit	0.392***	0.021	0.583***	0.193	0.425***	0.022	0.555***	0.205
Ecoupon	0.103***	0.002	0.059***	0.017	0.103***	0.002	0.060***	0.018
Wi-Fi	−0.104***	0.002	−0.061***	0.016	−0.093***	0.002	−0.057***	0.017
Demographic
Gender	−0.027***	0.002	−0.119***	0.023	−0.042***	0.003	−0.144***	0.024
Age	0.009***	0.000	0.017***	0.001	0.013***	0.000	0.021***	0.001
Dependents	0.038***	0.001	0.035***	0.006	0.034***	0.001	0.022***	0.007
Employment Status	Yes	Yes	Yes	Yes
Occupational Status	Yes	Yes	Yes	Yes
Time fixed effects	Yes	Yes	Yes	Yes
Observations	2,078,808	2,078,808	1,890,317	1,890,317
Log Likelihood	−3573963.7	−51492.2	−3243859.7	−46171.4
Pseudo${\mathit{R}}^{\mathbf{2}}$	0.0062	0.0081	0.0071	0.0086

Note: ***p < 0.01.

The results from the first stage of our proposed model (Equation (2)), which captures the drivers of consumer decisions regarding whether to watch a movie in the mall or not, are reported in Table 4 . The findings suggest that younger, male, and employed consumers are more likely to watch a movie in the mall than their counterparts. The probability of watching a movie also increased with the number of dependents and weekend visits. Finally, the number of movies each consumer watches per year, the instrument, is statistically significant.3

Table 4.

Results of the first-stage movie intention model.

Variables	Movie intention
Frequency	0.0243*** (0.0001)
Age	−0.0025*** (0.0003)
Gender (Reference = Female)	0.0241*** (0.0068)
Employment (Reference = Salaried)	−0.0442*** (0.0123)
Occupation (Reference = Officer)
Manager	0.0639 (0.0415)
Higher than manager	−0.0709 (0.2564)
Salary	5.25e-11 (3.62e-10)
Dependents	0.0104*** (0.0018)
Weekend	0.4415*** (0.0054)
Constant	−2.6434*** (0.0435)
Number of observations	2,360,192
Log Likelihood	−135441.1

Note: Standard errors are in parentheses; ***p < 0.01.

The parameter estimates associated with consumers' hedonic purchases conditional on movie-watching in the mall are reported in Table 5 . As shown, watching a movie does not increase hedonic (grocery) purchases at the supermarket (H₂ is not supported). Meanwhile, the effect of movie consumption on grocery shopping behavior among moviegoers varies depending on the movie genre consumers watch. Fantasy encourages moviegoers to engage more in hedonic purchases; hedonic products in general (${\beta }_{6,\mathrm{f}\mathrm{a}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{s}\mathrm{y}}=3.6869$; p < 0.01) and hedonic grocery products (${\beta }_{6,\mathrm{f}\mathrm{a}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{s}\mathrm{y}}=3.3820$; $p<0.01$). On the contrary, thrillers reduce moviegoers’ spending on hedonic products; hedonic products in general (${\beta }_{6,\mathrm{t}\mathrm{h}\mathrm{r}\mathrm{i}\mathrm{l}\mathrm{l}\mathrm{e}\mathrm{r}}=-0.2469$; p < 0.01) and hedonic grocery products (${\beta }_{6,\mathrm{t}\mathrm{h}\mathrm{r}\mathrm{i}\mathrm{l}\mathrm{l}\mathrm{e}\mathrm{r}}=-0.2519$; $p<0.01$), supporting H₃. Hedonic purchases at the supermarket are also significantly influenced by the number of empty seats when the consumer watches a movie. The emptier seats there are, the more hedonic purchases the moviegoers make after watching the movie (${\beta }_7=0.0006;p<0.05)$, supporting H₄.4 These findings highlight the role of both content and contextual factors associated with movie-watching in hedonic purchases.

Table 5.

Effect of the movie genre watched on subsequent hedonic purchases.

Variables	ln(Purchase amounts of hedonic products)		ln(Purchase amounts of hedonic groceries)
	Coef.	Std. err.	Coef.	Std. err.
After	0.0772	0.1506	0.1142	0.1548
MajorStudio	0.1337***	0.0371	0.1217***	0.0380
SeatsAvailable	0.0001	0.0002	−0.0001	0.0003
Quantity	0.0439***	0.0010	0.0342***	0.0101
After × MovieGenre (Ref = Bibliography)
Fantasy	3.6869***	1.1637	3.3820***	1.1443
Thriller	−0.2469**	0.1134	−0.2519**	0.1159
After × Quantity	0.0097	0.0133	0.0100	0.0136
After × SeatsAvailable	0.0006**	0.0002	0.0005*	0.0003
StoreVisited
LeisureEnt	0.0755***	0.0224	0.0734***	0.0231
Fashion	0.1961***	0.0534	0.1760***	0.0542
Segment
HotelVisit	0.3822	0.2374	0.3450	0.2496
Ecoupon	0.0344*	0.0206	0.0439**	0.0212
Wi-Fi	−0.0522***	0.0202	−0.0708***	0.0208
Inverse Mills Ratio	−0.2342***	0.0532	−0.2834***	0.0551
Movie genre fixed effects	Yes	Yes	Yes	Yes
Time fixed effects	Yes	Yes	Yes	Yes
Observations	2,342,896		2,341,755
Log Likelihood	−102276.1		−94499.9
Model Fit (${\mathit{\chi }}^{\mathbf{2}})$	357.41***		294.91***

Note: ***p < 0.01, **p < 0.05, and *p < 0.10.

For the control variables, we find that consumers who visit entertainment or leisure facilities (other than movie theaters) and clothing stores spend more on hedonic products and groceries than those who do not (${\beta }_8=0.0755;p<0.01$ and ${\beta }_9=0.1961;p<0.01$). We also find clear evidence of selection with respect to movie-watching in purchases at supermarkets (${\beta }_{IMR}=-0.2342;p<0.01)$. The negative coefficient of the IMR suggests that consumers who come to watch a movie have less incentive to purchase hedonic products in the shopping mall, supporting the findings from the estimation of Equation (1).

5. Discussion and conclusion

Retailers relying on face-to-face commerce have suffered from the COVID-19 crisis, as consumers have shifted their spending from in-store to online shopping amidst the pandemic. Given the dramatic changes in consumers’ spending habits, shopping malls have placed more emphasis on consumer experiences through entertainment facilities such as movie theaters, aiming to increase their foot traffic and cross-selling revenue per visit in the post-pandemic era. However, the amount of incremental revenue generated from such facilities remains unclear because consumers dedicated to the facilities are reluctant to spend at retail stores in shopping malls. We empirically tested the potential effect of watching movies on shopping mall retail revenue. This study adds to the literature on cross-selling and provides actionable implications for various retail operators, including shopping malls, which can help them survive in the new market environment post-pandemic.

5.1. Theoretical implications

The present study contributes to existing research in several ways. First, this study attempts to show the relationship between non-retail experience and hedonic consumption in shopping malls, advancing extant cross-selling literature. We explicitly model the relationship between movie-watching and retail purchases and show the potential negative effects of entertainment facilities on retail revenue in shopping malls. Extant literature has acknowledged the importance of retail-entertainment complexes in increasing retail revenue in shopping malls (Farrag et al., 2010; Ibraham and Ng, 2003), with the general expectation that consumers who go to the movies will stay longer and visit other stores on their trip. This is also consistent with current practice, where shopping malls provide a variety of entertainment options for consumers in addition to retail spaces. Contrary to conventional expectations, our findings suggest that such entertainment facilities may not be helpful for inducing cross-selling opportunities as expected when consumers visit the mall solely for entertainment experiences with no intention of spending money at retail stores.

Second, prior research has largely ignored the contextual factors of non-retail experience in cross-selling. Building on theoretical foundations that a feeling is known to affect future behaviors, we examined the relationship between content and contextual factors during non-retail experiences that might affect consumer feelings and cross-selling. Our results suggest that cross-selling is affected by content and contextual factors associated with the entertainment experience, such as movie genre and the number of empty seats. In particular, consumers who watch fantasy (thriller) movies increase (avoid) hedonic consumption during their subsequent shopping. Additionally, we find that consumers’ internal reactions and behavioral intentions may vary depending on the viewing environment inside the movie theater (Evans and Howard, 1973; Wakefield and Blodgett, 1999). The discomfort driven by a crowded environment can negatively affect the emotional state and reduce subsequent hedonic consumption in the shopping mall. These findings associated with the mechanism behind cross-selling bolster extant cross-selling literature which primarily focuses on the antecedents and consequences of cross-selling (Béal et al., 2019; Dahana et al., 2020; Evanschitzky et al., 2017; Reinartz et al., 2008).

5.2. Managerial implications

In addition to the theoretical contributions noted above, the findings of this study also have significant managerial implications. First, given that consumers tend to reduce their shopping consumption after watching a movie, shopping malls should be able to allocate marketing resources to cross-selling within entertainment facilities. Our data reveal that consumer spending on food and beverages in movie theaters contributes substantially to the overall retail revenue of the mall (8.76% of the total grocery spending from consumers who do not watch a movie). As such, encouraging consumers dedicated to entertainment in the mall to purchase relevant products within the movie theater could result in incremental revenue and mitigate potential losses incurred from operating entertainment facilities.

Second, our study offers managerial implications concerning the identification of effective targeted promotion strategies via demographic, category purchase, and product promotion information to maximize cross-selling revenue. In most retail settings, cost considerations limit firms from launching too many promotion campaigns, and thus allocating the campaign resources to which product types and consumer segments has been a primary interest (Saha et al., 2021). This study responds to this call, showing that consumers tend to adjust their hedonic purchases according to their psychological state after watching a movie. Essentially, our findings suggest that shopping malls can enhance the effectiveness of their targeted promotions by customizing them based on the content and contextual factors of movie watching. For example, they could provide coupons for relatively expensive hedonic products such as accessories and music equipment to consumers who come to watch a fantasy movie or visit the movie theater during less crowded movie showtimes. Such targeted promotion strategies should also be preferred to multi-product or service retailers associated with large ticket items, as they have observed a wave of “pent-up demand” followed by a remarkable post-pandemic recovery.

5.3. Limitations and recommendations for future research

Our study is not without limitations that may require further research. First, our empirical framework is calibrated only on consumer behavior from the pre-pandemic period. Given that the impact of entertainment experiences on consumers' emotional status and subsequent purchase behavior is not expected to change due to the pandemic, however, our findings provide useful managerial insights into the post-pandemic recovery. There are still empirical questions that need to be addressed, including whether the outbreak of COVID-19 changed shopping malls’ resource allocation to movie theaters and supermarkets, and whether such a decision impacted consumer behavior in shopping malls.

Second, it is important to note that the effect of movie-watching on cross-selling in our study should be viewed as descriptive rather than causal, given that our sample in the empirical study only considered customers who watched a movie. The causal effect of movie-watching on cross-selling can be achieved when we examine the situation where movie theaters do not exist. We leave this for future research.

Third, our findings could depend on potential mall-specific characteristics unobservable to the researcher, e.g., the location and accessibility of the movie theater. The availability of such data may allow us to develop boundary conditions for the effect of consumer intention to visit shopping malls. To further generalize our findings, it would be beneficial to include retail contexts beyond supermarkets in geographically diverse settings. Uncovering cross-selling opportunities driven by the psychological states associated with entertainment experiences is highly relevant to practice. In particular, targeted promotions and advertisements are important tools in dual-channel retailing (Kar et al., 2023). The findings of our study can be integrated into this research stream by proposing the potential determinants of hedonic buying behavior.

Declaration of competing interest

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability

The authors do not have permission to share data.