Engagement and Brand Recall in Software Developers: An Eye-Tracking Study on Advergames

Abstract

Advergames represent a novel product placement strategy that surpasses traditional advertising methods by fostering interaction between brands and their target audiences. This study investigates the unique engagement opportunities provided by video games, focusing mainly on the ‘flow experience’, an intensified state of immersion frequently encountered by players of computer games. Such immersive experiences have the potential to significantly influence a player’s perception, offering a new avenue for advertisements to impact and engage audiences effectively. The primary objective of this research was to examine the influence of advergames on players who are deeply immersed in the gaming experience, with a specific focus on the subsequent effects on brand recognition over time. The study involved 44 software developers, who were evenly divided into two groups for the experiment. Both groups were exposed to an identical gaming environment with the task of locating a designated product within the game. However, one group interacted with an enhanced version of the game, which included additional stimuli—such as dynamic music, an engaging narrative, time constraints, a competitive leaderboard, and immersive voice acting—to intensify the gaming experience. The experiment strategically placed various products within the game, and their detectability was assessed using eye-tracking technology. Following gameplay, participants completed questionnaires that measured their experience with flow state and brand recall. The data were analyzed using the Mann–Whitney U test and correlation analysis to facilitate comparisons. The findings indicated that the product associated with the primary task achieved the highest recall rate between both groups. Furthermore, eye-tracking technology identified the areas in the game that attracted the most attention, revealing a preference for mid- and high-level placements over lower-level ones.

1. Introduction

Games play an essential role in human life, acting as a social glue that unites individuals with shared interests. They even facilitate communication and interaction among people of diverse nationalities and cultures. In particular, gaming provides an opportunity to experience emotions and sensations that may be encountered in real life. Several motivating factors—such as competition, cooperation, socialization, discovery, and strategy [1]—emerge during gameplay, encouraging players to voluntarily engage with game elements, including music, storylines, time constraints, leaderboards [2], and voice acting [3]. These elements drive players to invest significant effort in mastering the rules of the game [4]. Rapid advances in computer technologies in recent years have catalyzed substantial growth in the gaming industry, leading to increased collaboration between gaming companies and other sectors. Research into the application of computer-based games has been carried out to develop solutions in various fields, including education, advertising, serious games, and politics [5]. Notably, the concept of advertising through gaming—primarily through ‘Adgames’, also known as ‘Advergames’, a form of ‘Advertising Games’—has gained prominence as an innovative strategy [6]. These games are specifically designed to promote a particular brand or advertising slogan [6], thus reaching a global audience far beyond the capabilities of traditional media [7]. Significantly, advergames can improve the connection between a brand and consumer perceptions [8], effectively convey company messages [9], increase brand awareness [10], rapidly convey the knowledge of marketing campaigns [11], and seamlessly integrate marketing content [12].

In the early 1980s, some brands began experimenting with in-game advertisements. One such pioneering game, “Tooth Protector”, was designed to disseminate messages about tooth protection and oral hygiene and was released for the Atari 2600 gaming console in 1983 [13]. In this game, players earned points by shielding teeth from snack attacks and preventing decay, thus delivering a social message about dental care. Advergames not only serve to engage users but also collect valuable data on potential consumers of the advertised products or brands, thereby effectively broadcasting their messages to the public. Examples include the “U.S. Army Game”, which aims to boost recruitment for the United States Army, and “Smokefree”, which educates players about the risks of smoking and its links to cancer, promoting a vital public health message [6].

Moreover, advergames have found their specialized area in racing video games, such as “Need for Speed: Porsche Unleashed”, which features Porsche automobiles [14], and “Gran Turismo”, where AUTOart included and later produced a vehicle bearing the Red Bull logo in real life after the game’s release [15]. However, it should be noted that some video games that are not primarily designed for advertising incorporate brand advertisements within their gameplay. An example is “SimCity”, where players can build cities from scratch and, in a subsequent version, could add the “Nissan Leaf Charging Station”, which showcases Nissan’s electric vehicles and positively influences happiness in the game of Sims for the virtual inhabitants of the game [16]. On December 6, 2018, Capcom announced the introduction of “Sponsored Content” in “Street Fighter V: Arcade Edition”, further illustrating the integration of advertising content within video games [17]. This feature allowed players to opt out of seeing advertisements (‘ads’ in the game) or to accept them in exchange for in-game currency. Capcom Cup Pro Tour 2018 logos were placed on the fighter gloves and costumes. However, on 17 December 2018, Street Fighter series producer Yoshinori Ono announced via Twitter that in-game advertisements would be removed from the Street Fighter game [18]. In the mobile game development industry, advertisements are commonly used in free games available on Google Play. For example, some advertisements were placed in the game “2048”, redirecting players to the advertisement homepage if they clicked on the ads, whether intentionally or not [19]. “Godzilla: King of the Monsters”, a movie released in 2019, featured Godzilla-themed visual items on PUBG Mobile t-shirts [20]. “Chipotle’s The Scarecrow”, a game released by a Colorado-based Mexican cuisine fast-food restaurant in 2013 [21,22], conveyed a message about protecting food and crops from harmful substances and using healthy ingredients. The associated short video became quite popular, having 13 million views and attracting 16,000 comments on YouTube [21]. Another advergame promoted healthy nutrition, encouraging children to make informed decisions about food [23]. The results showed that the children could differentiate between healthy and unhealthy foods [23]. The integration of gamification in advertising has become a significant focus in contemporary marketing strategies. Playing a game can positively affect brands, especially those initially unfamiliar, compared to brands without developed advergames [24]. When a brand is embedded in a game and the brand seems more exciting, gamers tend to develop a positive perception of this brand by making a connection between the attitude of their brand and the emotional response [25,26,27].

The integration of gamification in advertising has become a significant focus in contemporary marketing strategies. This technique leverages game mechanics in non-game contexts to enhance user engagement, brand recall, and consumer behavior. Advergames significantly affect brand recall and purchase intention [28]. Numerous studies have demonstrated the effectiveness of gamification in increasing brand awareness and consumer interaction. For example, Tian [29] found that advergame product placements significantly boost brand recognition and recall, making gamification a valuable tool for marketers. Similarly, van Berlo [25] conducted a meta-analysis that revealed advergames generally improve consumer attitudes towards brands. However, the impact on brand memory can vary, with younger consumers showing higher sensitivity to these persuasive effects. Additionally, the appropriate combination of game elements such as music, sound, characters, and visual effects helps to keep gamers interested in the advertised brand [30]. On the other hand, effective branding is crucial in the game because players’ attitudes towards the advertised brand are proportional to their gaming experience, positive or negative [31].

The psychological mechanisms of gamification have been a subject of extensive exploration in the literature. A theoretical model based on cognitive evaluation theory explains how gamification enhances user engagement by fulfilling psychological needs for autonomy, competence, and relatedness [32]. This model is strengthened by findings from Pour [33], who demonstrated that gamification positively influences customer experience through enhanced brand engagement, particularly in online retail contexts. Additionally, key gamification features—immersion, achievement, and social interaction—significantly enhance advertising effectiveness in social media environments [34].

The role of contextual factors in gamified advertising has been underscored in several studies. For instance, Sreejesh [35] examined how the platform and device used to access gamified content impact brand memory and attitude. Their research suggests that the interaction between these factors can significantly influence consumer engagement, with mobile devices being particularly effective in fostering deeper engagement and flow experiences. Similarly, the effects of gaming platforms and game speed on consumers’ memory and purchase intentions found that PC-based advergames generate better-delayed memory and influence purchase intentions more effectively than mobile-based games [36]. Furthermore, it has been noted that slower-paced advergames with famous brand placements lead to higher brand recall, particularly among users with more excellent persuasion knowledge [37].

Moreover, gamified advertising has shown significant potential in influencing consumer purchase intentions. Rialti [38] investigated how gamification in advertising can drive in-app purchases, emphasizing that experiences providing social, personal, and hedonic benefits can significantly enhance consumers’ intentions to purchase. Additionally, Vashisht and Sreejesh [39] explored how game speed and product harmony affect brand recall and attitudes, concluding that slow-paced games with low game–product harmony can result in higher brand recall. Larger brand sizes in advergames are more likely to be remembered and recognized by players, further supporting the importance of advergame design in enhancing brand memory [40].

The effectiveness of advergames has been found to vary based on several other factors, including brand fame, game involvement, and persuasion knowledge. For example, Vashisht and Pillai [41] reported that famous brand placements in advergames with high game involvement result in higher brand recall but less favorable brand attitudes. In other studies, such as those by Sukoco and Wu [42], advergames enhance consumer telepresence and affective responses, particularly for products with experience attributes. Moreover, the effectiveness of advergames in increasing children’s brand recognition and positive attitudes toward unhealthy food brands indicates a need for regulatory oversight [43]. These findings collectively underscore the complex interplay of factors that influence the effectiveness of gamified advertising strategies.

To understand the motivations of the players, the concept of flow, which refers to the optimal experience characterized by complete immersion in an activity, resulting in a deep sense of satisfaction and happiness while playing [44], is measured and analyzed in certain games [45,46,47]. When players experience flow, their concentration on the game increases, as does the duration of their gameplay, which corresponds to the depth of their experience [46]. A relationship has been established between the visual attention of game players, the flow experience, and game-based learning [47]. Furthermore, it is hypothesized that game players exhibit improved cognitive learning in the flow state [48]. Furthermore, a significant relationship has been identified between flow and how advergames influence player buying behaviors, with flow experiences that facilitate this connection [8]. There is a link between playing a game and Game-Based Learning (GBL) because players are required to make rapid decisions within limited time frames, listen to game messages or visual goals, and design optimal strategies [48]. The results indicate that the players achieve better conceptual understanding in GBL when in the flow state [48]. Flow is characterized by nine main components: Challenge–Skill Balance, Action–Awareness Merging, Clear Goals, Feedback, Concentration on Task, Control, Loss of Self-consciousness, Transformation of Time, and Autotelic Experience. Based on these flow components, the cognitive states of the players can be described in terms of how they feel when focusing on a game-based activity [49].

Human–computer interaction (HCI) examines the processes of the human brain, focusing on how humans interact with systems, and furthering system development based on established learning principles. Eye-tracking, a form of HCI, functions to collect data on individuals’ eye movements [50]. This technology is utilized to visualize where players are looking on a screen, using various colors or intensity indicators to represent eye movements [51]. Eye tracking is instrumental in understanding human behavior in various areas [50]. Data derived from the movements of the eye of a player can elucidate usability and identify issues related to the game [52]. Although surveys or interviews subjectively assess cognition, eye-tracking technology provides an objective evaluation, providing information on eye movements, screen captures, and task completion times [47]. Eye-tracking measures, such as the percentage of duration in a zone and the percentage of fixation count within the zone, reveal the relationship between flow and time distortion [48]. Eye-tracking methods are used to analyze human comprehension activities, where the duration of fixation is indicative of problem-solving processes in game-based learning (GBL) [53]. Fixations are reflective of human comprehension processes [54], with higher values suggesting more significant cognitive effort [55]. Furthermore, eye-tracking devices have been adapted for innovative applications, such as serving as input devices or game controllers in gaming applications [56]. Several metrics of fixation, such as areas of interest (AOIs) [57], gaze duration, fixation count [58], fixation duration [59], visit count, visit duration [60], and heatmaps [61], are employed in this context.

2. Materials and Methods

The research questions focused on the effectiveness of advergames, the relationship between the flow experience and various game elements, and the extent of players’ engagement with the game screen. For the study, two games that share the same concept but incorporate different aspects of the game were developed. The design process began with the creation of prototypes on paper, which were later evaluated. The development phase constituted the third iteration. Upon the completion of these iterations, the advergames were finalized, and the study’s participants then proceeded to test the games.

Data were collected from brand-and-product matching, flow questionnaires, players’ comments, and eye-tracking results.

Table 1. Research design matrix adapted from [47].

Hypothesis	Data Source	Method
H1: Players can remember task-related brands of the products.	Matching of brands and products	Mann–Whitney U test
H2: Players experience flow when the number of game elements increases.	Flow State Scale Players’ comments	Mann–Whitney U test Correlation analysis
H3: Players focus more on the screen location corresponding to their eye level.	Eye-tracking data	Heatmap analysis AOI analysis

provides information about the data sources and analysis methods used for each hypothesis.

For this research, two different games were created: one containing more game elements, while the other had fewer, such as the exclusion of a storyline, leaderboard, and time limit. The primary objective was to assess the impact of these elements of the game on the gaming experience of the players. Six brand names of task-related products were incorporated into the game, along with 12 other brands placed alongside the brands related to tasks. After playing the game, players were asked to match the products with their corresponding brands.

Jackson and Marsh [62] developed a questionnaire to measure the flow experience of athletes. For the current study, this questionnaire was adapted to the Turkish context, and its Turkish language version was administered to players immediately after their gameplay. The flow state scale (FSS) consists of 36 items in 9 dimensions, using a 5-point Likert scale (with four items per dimension) to measure the flow experience [62].

An additional question (#37) described the flow experience [63], asking players if they had directly experienced flow. The responses were recorded using a 5-point Likert-type scale. Two further questions (#38 and #39) inquired about factors that motivated and demotivated the players, respectively.

Players who identified with the given description of flow were asked to specify which factors motivated them. Conversely, those who did not identify with the flow description were asked to indicate the factors that demotivated them.

2.1. Game Design

The game is categorized as an “illustrative advergame” [64], establishing a direct association between the player and the brand. Designed with a supermarket theme, the advergame features specific products associated with brands, each represented by animal names and displayed on the virtual supermarket’s shelves. Players receive a shopping list instructing them to locate certain products. The brands chosen for these products are listed on the shopping list, while the alternative brands of the same products, not listed, are positioned on the top or bottom shelves. In particular, the task-related brand, “Lion Detergent”, is mentioned on the shopping list (located in the bottom corner of the screen), with its products strategically placed on the shelves. “Goat Detergent” is positioned on the top shelf, “Penguin Detergent” is positioned on the middle shelf, and “Lion Detergent” is positioned on the bottom shelf.

The game exclusively advertises “Lion Detergent”, making it a task-related brand, while “Goat Detergent” and “Penguin Detergent” do not correspond to any specific task. Players are expected to select the corresponding product when they notice the detergent name on their shopping list. The game ends once all items on the shopping list have been found.

Figure 1 illustrates a segment of the gameplay. It shows the shopping list in the bottom left corner of the screen with the “Lion Detergent” listed, guiding the player to locate the product on the supermarket shelves. The “Goat” and “Penguin Detergents” are also placed in proximity to the “Lion Detergent”.

2.2. Formal Game Elements

Two games based on the same concept were designed for this research study. However, certain gaming elements were omitted from the second game.

Player Interaction Pattern: The game involves a single player who assumes the role of a customer visiting a supermarket.

Objectives: The objective is to locate and collect all of the products listed on the shopping list within the supermarket.

Rules:

• Players enter the game by typing their name in the text box and clicking the “Add” button.
• Returning players should type their name and click the “Continue” button to proceed.
• Subsequent levels are unlocked only after completing the preceding level.
• The game is structured into three levels, each featuring different product placements.
• Level One tasks the player with finding one listed product within 120 s.
• Level Two requires finding two listed products within 110 s.
• Level Three challenges the player to find three listed products within 100 s.
• Players may retry a level if they fail to complete it within the allotted time.

Procedures: The game begins at Level One. Players must act swiftly to complete each level within the designated time.

Boundaries: The virtual supermarket comprises 12 shelving units stocked with various generic products. Only one shelving unit, containing six shelves, holds the products related to the task.

Controls: The game uses standard first-person gameplay controls. Players navigate their characters using the “W”, “A”, “S”, and “D” keys for movement and the mouse to change the viewpoint. To add an item from the shopping list to the basket, players click the mouse’s left button.

Table 2. Keyboard inputs and actions.

Controller	Input	Action
Keyboard	w	Move forward
Keyboard	a	Move left
Keyboard	s	Move backward
Keyboard	d	Move right
Mouse	Cursor movement	Look at location
Mouse	Left click	Add product to basket

provides a detailed breakdown of the keyboard and mouse controls.

Resources: The game features a single supermarket setting with 12 shelving units. A shopping list is displayed in the lower-left corner of the screen, held by a hand. The game includes various products and an elderly woman character who seeks help from the players. In addition, a leaderboard displays the players’ scores. Game mechanics also include a timer, ambient background sounds, a warning sound for the final 20 s of gameplay, and an audio cue that signals the completion of a task from the shopping list.

Figure 2 introduces Game 1, where the elderly woman character says (translated), “My dear, I’m too old and tired to shop. Could you help me pick up a few things from the supermarket?” Following this interaction, the player enters their name, thus activating Level One of the first game.

Task-Related Products: “Lion Detergent”, “Chick Milk”, “Rooster Cheese”, “Zebra Wafer”, “Deer Tea”, and “Panda Chocolate”.

Non-Task-Related Products: “Elephant Cheese”, “Giraffe Cheese”, “Goat Detergent”, “Penguin Detergent”, “Butterfly Chocolate”, “Rabbit Chocolate”, “Bird Wafer”, “Chicken Wafer”, “Pelican Milk”, “Squirrel Milk”, “Fox Tea”, and “Owl Tea”.

2.3. Dramatic and Dynamic Elements

Premise: At the start of the game, a weary elderly woman requests the player’s assistance in purchasing items from her shopping list at the supermarket. To enhance the game’s dramatic effect, the elderly woman’s character is brought to life not only through visual representation and textual description but also by incorporating her voice into the game.

Character: The game employs a first-person perspective [65], which means that the graphical viewpoint is rendered from the player’s perspective. In this mode, players do not see an avatar body within the game; instead, they view the virtual environment as if through their own eyes.

Various challenges were incorporated into the game to ensure a balanced gameplay experience. The difficulty increases progressively with each level, beginning with Level One as the easiest. At Level One, players are required to collect one item from the shopping list, while at Level Three, the task involves collecting three items. Additionally, the allotted completion time decreases with each advancing level.

Excluded Game Elements for the Second Game: A second game was developed, omitting certain elements that were present in the first game. The elderly woman character, the leaderboard, and the timer were all removed. In this version, players can continue playing a selected level until all products are found, with no time constraints. The levels were renamed ‘Shopping List One’, ‘Shopping List Two’, and ‘Shopping List Three’. Additionally, the background sound, the warning sound for the remaining time, and the sound indicating the completion of tasks from the shopping list were also eliminated from the second game. Unlike in the first game, where the shopping list was depicted as being held in the player’s (virtual) hand, in the second game, the shopping list is placed inside a box. Apart from these modifications, all other aspects of the second game remain consistent with the first. Figure 3 shows the introduction of the second game.

2.4. Implementation

The game development process concluded after the third iteration. In the initial iteration, the supermarket model was integrated into Unity, where the lighting and the supermarket’s perspective were fine-tuned. Subsequently, a first-person character was designed, endowed with functionalities that allowed movement across the stone flooring based on user keyboard inputs (W, A, S, and D) and perspective changes via mouse control. During the second iteration, brands and their respective products were crafted in Blender and then positioned on the supermarket shelves. The third iteration saw the addition of gaming elements. To ensure game balance, new elements were introduced and play-tested after each iteration, with adjustments made based on play-test feedback. This iterative process continued until the game’s rules were fully established. The development of the advergame was considered complete following thorough play-testing and finalization of the prototype, with the entire implementation phase concluding within a month.

The Supermarket 3D model was acquired from https://assetstore.unity.com (accessed on 9 May 2019). The 3D models for both task-related and non-task-related brands were crafted using Blender (free version 2.79), while the graphics were created in GIMP (version 2.10.10). Each brand, assigned an animal name, was labeled in Turkish. Images of these animals were sourced from Google, ensuring they were marked ‘free to use or share’ under usage rights. Figure 4 showcases one of the brand models developed in Blender.

The Unity 3D game engine, version 2017.1.0f3, commonly utilized by students, served as the development platform for both games, employing Unity C# scripts for game creation. Figure 5 depicts the 3D game development environment within Unity, while Figure 6 presents the class diagram. The game classes inherit from the Unity MonoBehaviour base class to facilitate interaction between the player and UI components. Bitbucket was used for code version control.

2.5. Experimental Design

Eye-tracking analysis plays a crucial role in human–computer interaction (HCI) due to its ability to digitize individuals’ eye movements and compile comprehensive data on precisely where users are looking on a screen. By employing such analysis, computer interfaces can be optimized to be more intuitive and efficient, facilitating quicker and easier access to critical information. In this study, eye-tracking analysis was used to identify the screen locations that captured the most attention from players during gameplay. The analysis focused on several key metrics: fixation, the area of interest, gaze duration, fixation count, fixation duration, visit count, and visit duration.

The Tobii X2-60 compact eye tracker was leased from Bilten Bilişim for several weeks. The first week was dedicated to familiarizing with the eye tracker’s functionalities and mastering its practical use. The subsequent two weeks were devoted to collecting eye-tracking data, during which time the device served as the primary tool for recording real-time eye movements.

A magnetic holder was attached to the bottom of the screen frame to secure the eye tracker. The optimal viewing distance between the participant’s eyes and the eye-tracking device is recommended to be approximately 65 cm [66]. Figure 7 illustrates the setup, showing the distance between the participant’s eyes and the eye-tracking device.

Before starting data collection, each participant underwent an eye calibration process to account for individual differences in the precise location of the fovea. This calibration process is crucial to ensure that the eye tracker is accurately aligned with the eye characteristics of each participant, thus tailoring the gaze point calculations to individual specifications [59]. Figure 8 shows a successful calibration of the eye tracker device.

The test games were run on an ASUS 5500U laptop with the following specifications: Intel Core i7 Processor, 2.40 GHz

• 1366 × 768. Screen Resolution
• 15.6. -inch Display Monitor
• 8. GB RAM, 64-bit
• Windows 8.1 OS

2.6. Experimental Test

Within this research, two games were developed. The first game was designed with various enriched gaming elements, whereas the second game was intentionally simplified by omitting several of these elements. Throughout the gameplay, participants’ eye movements were captured by the eye-tracking device. An office space was designated for conducting the tests, where measures were taken to ensure an optimal testing environment: window blinds were closed, the office door was closed, and all mobile phones were off to minimize potential distractions. Each session involved a single participant (player) and the researcher. Before beginning a game, participants were briefed about eye-tracking technology and gameplay. Figure 9 shows a participant participating in one of the games.

In HCI studies, it is considered adequate to have fewer than 30 participants [67], with recommendations suggesting that quantitative results can be effectively obtained from 20 participants [68]. For comparing two groups, having 13 to 50 participants in each group is deemed optimal when applying eye-tracking analysis [69]. To explore the relationship between flow and gameplay, 22 participants were used [48]. In this study, two groups were established, each comprising 22 participants. The first group was designated as the experimental group, and the second as the control group. Consequently, a total of 44 software developers volunteered to participate in the advergame test, during which an eye-tracking device recorded their eye movements.

Before the test began, potential participants were evaluated for their video game experience and subsequently divided into two groups based on this criterion. They were asked to rate their previous video-gaming experience on a scale from 0 to 100. Those who indicated “0” as their level of prior video-gaming experience were excluded from participation. Selection criteria required a minimum video game experience score of 10 and a maximum of 100. In forming the groups, participants with similar levels of video game experience were allocated to different groups. For example, if two participants each had a score of 60, they were placed in separate groups. A difference of 10 points on the scale was considered acceptable to categorize participants as having similar levels of experience. All participants were experienced software developers with varying degrees of prior video game experience.

Table 3. Dependent and independent variables.

Hypothesis	Dependent Variables	Independent Variables
H1: Players can remember task-related brands of the products.	Number of remembered brands and products after playing an advergame.	Task-related brands in the advergame.
H2: Players experience flow when the number of game elements increases.	Players flow measurements.	Game elements.
H3: Players focus more on the screen location corresponding to their eye level.	Size of player’s gaze points.	Player’s eye levels on the corresponding positions.

delineates the dependent and independent variables associated with each hypothesis. For the first hypothesis (H1), the dependent variables are the number of brands and the products recalled, while the promoted brands constitute the independent variables. Regarding the second hypothesis (H2), the flow measurement serves as the dependent variable, with the game elements acting as the independent variables. For the third hypothesis (H3), the aggregated gaze points represent the dependent variables, whereas the players’ eye level at the corresponding screen position is identified as the independent variable.

3. Results

The experimental and control groups were established for the study. The experimental group was involved with the first game, which was replete with numerous gaming elements. In contrast, the control group was involved in the second game, from which several gaming elements present in the first game were omitted. Both games featured identical brands and products within the virtual supermarket, and their placement on the shelves was consistent across both versions. During gameplay, the eye movements of the participants were tracked and recorded by the eye-tracking device. Following each gaming session, participants filled out questionnaires that captured their demographic information, assessed their flow experience, and inquired about the brands and products they recalled. The following sections detail the results, including analyses and evaluations. Descriptive statistics were compiled using Microsoft Office Professional Excel 2016, and statistical analyses were performed using IBM’s SPSS Statistics Package (Version 22).

In both study groups and across both game versions, a total of six promoted and twelve non-promoted brands were used, consistent for all groups and games. In addition, identical shelf positions were maintained in both versions of the virtual supermarket.

Participants were tasked with finding products listed on shopping lists within a virtual supermarket setting. For the first player group, corresponding to the first game, the levels were designated as ‘Level 1’, ‘Level 2’, and ‘Level 3’. Conversely, for the second player group associated with the second game, the levels were renamed ‘Shopping List 1’, ‘Shopping List 2’, and ‘Shopping List 3’. Although both groups were required to complete each level to progress successfully, the second group was not subjected to any warning of time limitation.

On completion, both groups completed all three tasks. The completion times, measured in seconds, are detailed in

Table 4. Statistical calculations and descriptive results of completion time.

	Task	N	Min	Max	Mean	SD
Group 1	Task 1	22	30	890	246	222
	Task 2	22	80	613	275	166
	Task 3	22	60	703	250	193
	Sum	22	253	2199	771	442
Group 2	Task 1	22	27	787	239	200
	Task 2	22	84	990	344	260
	Task 3	22	78	486	205	116
	Sum	22	242	1539	789	349

N: number of participants, Min: minimum completion time, Max: maximum completion time, Mean: mean completion time, SD: standard deviation.

. Group 1 (the experimental group) accumulated a total gaming time of 16,958 s (approximately 4.71 h), whereas Group 2 (the control group) totaled 17,353 s (approximately 4.82 h).

After completing the gameplay, players were required to match both the promoted and non-promoted brands with their respective products. This task served as the basis for testing the first hypothesis, with the results of the analysis presented in

Table 5. Mann–Whitney U Test Results.

	Brands	Min	Max	Median	Mean	SD	U	z	p
G1	Task-Related Brands	1	6	3	3.18	1.47	28.50	−5.15	0.00 *
	Non-Task-Related Brands	0	4	0	0.50	0.96
G2	Task-Related Brands	0	5	3	2.91	1.27	32.00	−5.14	0.00 *
	Non-Task-Related Brands	0	3	0	0.41	0.80

* p < 0.05, G1 = Group 1, G2 = Group 2. Min: minimum number of remembered brands, Max: maximum number of remembered brands, Mean: mean of remembered brands, SD: standard deviation, U: U value, z: z value, p: p value.

.

H_0: There is no significant difference between the two distributions.

H_1: There is a significant difference between the two distributions.

For Group 1, the statistical analysis of the remembered promoted brands revealed a median of three, a mean of 3.18, a standard deviation of 1.47, a highest value of six, and a lowest value of one. In contrast, for the non-promoted brands that were recalled, the statistics indicated a median of zero, a mean of 0.50, a standard deviation of 0.96, a highest value of four, and a lowest value of zero.

For Group 2, the statistical results for the remembered promoted brands displayed a median of three, a mean of 2.91, a standard deviation of 1.27, a highest value of five, and a lowest value of zero. The statistical results for the recalled non-promoted brands showed a median of zero, a mean of 0.41, a standard deviation of 0.80, a highest value of three, and a lowest value of zero.

The Mann–Whitney U test revealed a significant difference between the remembered promoted and non-promoted brands (U = 28.50, z = −5.15, p < 0.001) for Group 1. A similar significant difference was observed between the recalled promoted and non-promoted brands for Group 2 (U = 32.00, z = −5.14, p < 0.001), with the p-values being less than 0.05 in both cases.

Consequently, the null hypothesis (H_0) was rejected for both groups, as participants remembered a more significant number of promoted brands and their products compared to non-promoted ones after engaging with the developed advergame.

The decision for Hypothesis 1 is documented in

Table 6. Decision for Hypothesis 1.

H1: Players can remember task-related brands of the products.

Hypothesis is accepted.

, where the first hypothesis was confirmed.

3.1. Results of the Flow Measurement

The Flow State Scale comprised 36 questions, which were administered to the players after their respective game sessions. Both groups of players completed the questionnaire, which was categorized into nine distinct flow dimensions [62].

Table 7. Scale questions divided into nine dimensions [62].

Dimension	Questions
Challenge–Skill Balance	Q1	Q10	Q19	Q28
Action–Awareness Merging	Q2	Q11	Q20	Q29
Clear Goals	Q3	Q12	Q21	Q30
Feedback	Q4	Q13	Q22	Q31
Concentration on Task	Q5	Q14	Q23	Q32
Control	Q6	Q15	Q24	Q33
Loss of Self-consciousness	Q7	Q16	Q25	Q34
Transformation of Time	Q8	Q17	Q26	Q35
Autotelic Experience	Q9	Q18	Q27	Q36

enumerates each flow dimension along with the corresponding question numbers, with each quartet of questions corresponding to a specific dimension of flow. In addition to these structured questions, players were invited to respond to an open-ended question on the factors that motivated or demotivated them during the game. The analysis sought to determine whether there was a significant difference in flow measurements between the two study groups. To assess the reliability of each dimension, the Cronbach alpha values (α) were calculated. The Mann–Whitney U test was used to facilitate comparison between groups. This procedure was instrumental in evaluating the second hypothesis.

Table 8. Mean, standard deviation, and reliability analysis.

		Group 1			Group 2
Dimension	N	Mean	SD	a	Mean	SD	a
Challenge–Skill Balance	22	13.95	3.86	0.84	14.50	2.82	0.72
Action–awareness merging	22	12.82	3.03	0.56	12.86	2.32	0.25
Clear goals	22	15.59	2.67	0.57	16.50	2.54	0.74
Feedback	22	12.86	3.78	0.90	13.50	3.57	0.88
Concentration on task	22	15.55	2.63	0.58	14.95	3.55	0.71
Control	22	16.50	2.26	0.79	16.14	3.23	0.85
Loss of self-consciousness	22	15.23	3.50	0.74	15.18	3.51	0.85
Transformation of time	22	12.91	2.74	0.36	12.14	2.85	0.58
Autotelic experience	22	16.73	2.51	0.79	16.00	2.69	0.74
Flow State Scale	22	132.14	17.19	0.89	131.73	21.22	0.94

N: number of participants, Mean: mean of questions, SD: standard deviation, a: a value.

outlines the details pertaining to the flow dimensions and reliability analysis. The ranges of Cronbach’s alpha values were examined across all nine dimensions. The alpha coefficient and levels of internal consistency are as follows [70].

• α ≥ 0.9 Excellent
• 0.9 > α ≥ 0.8 Good
• 0.8 > α ≥ 0.7 Acceptable
• 0.7 > α ≥ 0.6 Questionable
• 0.6 > α ≥ 0.5 Poor
• 0.5 > α Unacceptable

An alpha value accepted as reliable is, therefore, 0.7 or above in this analysis.

Flow experience, as defined by Csikszentmihalyi [44], refers to a state of complete immersion and optimal engagement in an activity in which individuals lose track of time and self-consciousness. This psychological state is characterized by intense focus, a sense of control, and intrinsic enjoyment, making it an ideal construct to study in the context of interactive and engaging activities such as games. To assess the flow experience of participants during gameplay, we employed a well-validated flow questionnaire specifically designed to measure the multidimensional nature of the flow. The questionnaire comprises several items that evaluate key dimensions of flow, including concentration on the task at hand (e.g., “I was completely focused on what I was doing”), a sense of control over the activity (e.g., “I felt in control of the game”), the merging of action and awareness (e.g., “I was not aware of myself as separate from the game”), a loss of self-consciousness (e.g., “I was not concerned with how I was perceived by others”), and an altered sense of time (e.g., “Time seemed to pass quickly while playing”). Participants rated each item on a seven-point Likert scale ranging from one (strongly disagree) to seven (strongly agree). The questionnaire has been established in previous studies to have strong psychometric properties, demonstrating acceptable (Cronbach’s alpha > 0.7) and high internal consistency (Cronbach’s alpha > 0.8) and construct validity through factor analyses that confirm the dimensionality of the flow construct. In our study, the questionnaire also showed good reliability, with a Cronbach alpha of 0.85, indicating that the measure was consistent and reliable in capturing the flow experience of the participants during the advergame sessions.

The Challenge–Skill Balance dimension exhibited good reliability for Group 1 (α = 0.84) and acceptable reliability for Group 2 (α = 0.72). Conversely, the Action–Awareness Merging dimension displayed poor reliability for Group 1 (α = 0.56) and unacceptable reliability for Group 2 (α = 0.25). The Clear Goal dimension showed poor reliability for Group 1 (α = 0.57) but acceptable reliability for Group 2 (α = 0.74). The Feedback dimension demonstrated excellent reliability for Group 1 (α = 0.90) and good reliability for Group 2 (α = 0.88). The Concentration on Task dimension had poor reliability for Group 1 (α = 0.58) and acceptable reliability for Group 2 (α = 0.71). The Control dimension revealed acceptable reliability for Group 1 (α = 0.79) and good reliability for Group 2 (α = 0.85). The Loss of Self-consciousness dimension achieved acceptable reliability for Group 1 (α = 0.74) and good reliability for Group 2 (α = 0.84). The Transformation of Time dimension presented unacceptable reliability for Group 1 (α = 0.36) and poor reliability for Group 2 (α = 0.58). The Autotelic Experience dimension attained acceptable reliability for both Group 1 (α = 0.79) and Group 2 (α = 0.74). Overall, the Flow State Scale showed good reliability for Group 1 (α = 0.84) and excellent reliability for Group 2 (α = 0.94). The lower Challenge–Skill Balance dimension for Group 2 compared to Group 1 is attributed to the exclusion of several gaming elements. The Action–Awareness Merging dimension was considered unacceptable for both groups. The Clear Goals dimension for Group 1 was not only unacceptable but also lower than that for Group 2. The Feedback dimension scored higher for Group 1 than for Group 2, potentially influenced by the remaining time warning sound and task completion sound in Group 1. The Concentration on Task dimension was deemed unacceptable for Group 1 and was lower than that for Group 2, possibly affected by the removal of time limitation elements in Game 2. The Control and Loss of Self-consciousness dimensions were lower for Group 1 compared to Group 2. The Transformation of Time dimension was unacceptable for both groups. The Autotelic Experience dimension scored higher for Group 1 than for Group 2. Despite certain dimensions not meeting the acceptable standards, the overall Flow State Scale results indicated that both groups experienced flow. Cronbach’s alpha values greater than 0.7 were analyzed for each flow dimension, complemented by Pearson correlation analysis, as detailed in

Table 9. Pearson correlation analysis between flow dimensions for Group 1.

	CSB	F	C	LSC	AE
CSB	-
F	0.789 *	-
C	0.297	0.537 **	-
LSC	0.286	0.376	0.399	-
AE	0.490 *	0.442 *	0.654 **	0.197	-

**: Correlation is significant at the 0.01 level (two-tailed). *: Correlation is significant at the 0.05 level (two-tailed). CSB: Challenge–Skill Balance, F: Feedback, C: Control, LSC: Loss of Self-consciousness, AE: Autotelic Experience.

(for Group 1) and

Table 10. Pearson correlation analysis between flow dimensions for Group 2.

	CSB	CG	F	C	CT	LSC	AE
CSB	-
CG	0.787 **	-
F	0.782 **	0.838 **	-
C	0.828 **	0.909 **	0.862 **	-
CT	0.515 *	0.683 **	0.682 **	0.682 **	-
LSC	0.806 **	0.864 **	0.847 **	0.825 **	0.687 **	-
AE	0.639 **	0.732 **	0.630 **	0.686 **	0.528 *	0.635 **	-

**: Correlation is significant at the 0.01 level (two-tailed). *: Correlation is significant at the 0.05 level (2-tailed). CSB: Challenge–Skill Balance, F: Feedback, C: Control, LSC: Loss of Self-consciousness, AE: Autotelic Experience.

(for Group 2).

Table 9 shows the results of the correlation analysis for Group 1. It was observed that the Challenge–Skill Balance (CSB) dimension exhibits a weak relationship with both the Feedback (F) and Autotelic Experience (AE) dimensions. The Feedback dimension, however, demonstrates a strong relationship with the Control (C) dimension and a weak relationship with the Autotelic Experience dimension. Furthermore, the Control dimension is strongly related to the Autotelic Experience dimension.

Table 10 details the correlation analysis findings for Group 2. The Challenge–Skill Balance (CSB) dimension was found to have a strong relationship with the Clear Goals, Feedback, Control, Loss of Self-consciousness, and Autotelic Experience dimensions, as well as a weak relationship with the Concentration on Task dimension. The Clear Goals dimension demonstrated a strong relationship with the Feedback, Control, Concentration on Task, Loss of Self-consciousness, and Autotelic Experience dimensions. Similarly, the Feedback dimension exhibited a strong relationship with the Control, Concentration on Task, Loss of Self-consciousness, and Autotelic Experience dimensions. The Control dimension was strongly related to the Concentration on Task, Loss of Self-consciousness, and Autotelic Experience dimensions. The Concentration on Task dimension had a strong relationship with the Loss of Self-consciousness dimension but a weak relationship with the Autotelic Experience dimension. Furthermore, the Loss of Self-consciousness dimension showed a strong relationship with the Autotelic Experience dimension. It was observed that both groups exhibited a strong relationship between the Control and Feedback dimensions, as well as between the Control and Autotelic Experience dimensions. Mann–Whitney U tests were applied to the Flow State Scale results to determine whether there were significant differences between the two groups.

H₀: 

There was no significant difference between the two distributions.

H₁: 

There was a significant difference between the two distributions.

Table 11. Mann–Whitney U test for the Flow State Scale.

Dimension	Group	Min	Max	Median	Mean
Flow State Scale	G1	96	165	132.5	132.14
	G2	95	164	131.5	131.73
Dimension	Group	SD	U	z	p
Flow State Scale	G1	17.19	236.5	−0.129	0.897
	G2	21.22

Min: minimum weight of Likert scale, Max: maximum weight of Likert scale, Mean: mean of Likert scale, SD: standard deviation, U: U value, z: z value, p: p value.

displays the outcomes of the Mann–Whitney U test. As indicated in Table 11, no significant difference was observed between the Flow State Scale results of the two groups, as evidenced by a p-value exceeding the 0.05 significance level. Consequently, the null hypothesis (H₀) was accepted. The decision for Hypothesis 2 is documented in

Table 12. Decision for Hypothesis 2.

H2: Players experience flow when the number of game elements increases.

Hypothesis is rejected.

. Although both groups experienced flow, the correlations among the flow dimensions differed, leading to the rejection of the second hypothesis.

Based on the players’ feedback, the elements that motivated Group 1 included time limitations, visual elements, sounds, and the task of finding products. For Group 2, the motivating factors were the search for products, the immersive experience of feeling as though in an actual market, visual elements, and the player being the main character. However, the absence of sounds demotivated one player from Group 2.

3.2. Results of Eye-Tracking Analysis

The eye-tracking data were analyzed using Tobii Studio (version 3.4.8). Two separate projects were established: one for the Experimental Group and another for the Control Group. A total of 22 volunteers from the Experimental Group participated in the first game, while 22 volunteers from the Control Group participated in the second game. Eye movements were visualized as points in video recordings. Specific product scenes were segmented from the overall recordings. Subsequently, all eye movement data were incorporated into the selected scene, which encompassed the duration from the initial fixation within an area to the first fixation beyond it. Six distinct areas were identified, showcasing products such as detergent, milk, cheese, wafers, tea, and chocolate. Figure 10 illustrates the selection of the milk product captured from a segment of one of the recordings.

3.3. Heatmap Analysis

The heatmap results were derived from data based on the aggregated gaze points within the areas of interest (AOI). The analysis generated a total of six heatmaps. The settings for these heatmaps included a radius of 87 pixels for color representation, with the data type set to ‘count’ and a zoom level fixed at 44%. The color scheme ranged from intensive to sparse points, transitioning from red to yellow to green.

Table 13. Heatmap analysis of areas of interest.

Group 1	Group 2
Detergent	Detergent
Milk	Milk
Wafer	Wafer
Chocolate	Chocolate
Cheese	Cheese
Tea	Tea

outlines the heatmap results for both study groups. It is important to note that the branded products were randomly placed on various shelf positions.

The red areas on the heatmap indicated the regions that captured the most attention from the players, signifying where their focus was most intense. Heatmap analysis revealed that players typically directed their gaze at eye level, even when target products were placed on the bottom shelf. Notably, ‘Lion Detergent’ and ‘Panda Chocolate’ were listed on the shopping list but placed on the bottom shelf. However, players tended to concentrate on ‘Penguin Detergent’ and ‘Butterfly Chocolate’ instead.

3.4. Area of Interest (AOI) Fixation Count Analysis

While attempting to locate specific products listed on the shopping list, players examined the positions of all products. Fixation counts represent the total number of data points collected from an area of interest (AOI) within a scene.

As indicated in Table 13, the products placed on the middle and top shelves attracted the most attention from the players.

Table 14. Decision for Hypothesis 3.

H3: Players focus more on the screen location corresponding to their eye level.

Hypothesis was accepted.

documents the decision regarding Hypothesis 3, which confirms the acceptance of the third hypothesis.

4. Discussion

The current study aimed to evaluate the effectiveness of advergames, the flow experience, and the visual attention of game players. It was found that players could recall brands or products that were promoted within an advergame but struggled to recall non-promoted brands or products after gameplay. This finding suggested that for an advergame to be effective, the advertised brand or product needed to engage players on some level. However, the study also noted that players were not always interested in the brands or products displayed in games. Furthermore, the results supported the view that players tended to have difficulty remembering brands placed on billboards within games [71].

A second significant finding was that the removal of certain gaming elements did not prevent players from experiencing a state of flow in advergames. Both groups of players in this study experienced flow, albeit through different dimensions. Notably, the dimensions of Feedback and Control demonstrated a strong relationship for both groups, indicating that clear feedback enabled players to maintain a sense of agency within the game. Similarly, the dimensions of Control and Autotelic Experience were strongly correlated, suggesting that the ability to act freely within a game fostered intrinsic motivation to engage in the activity. The influence of other dimensions varied depending on the presence or absence of specific gaming elements. The Challenge–Skill Balance dimension, which typically reflected the game’s difficulty level, was often a motivating factor for players. However, excessive difficulty could negatively impact the player’s experience through other dimensions.

The insights from our research shed light on the impact of advergames on brand recall and attention within the specific context of our designed games. It is important to acknowledge that these findings may not be universally applicable to all game types. The simplicity and controlled environment of our study allowed us to analyze specific variables. However, more complex games with different mechanics, narratives, and visual designs could yield different outcomes. Individual differences such as sitting posture, height, and ergonomic factors could have influenced the eye-tracking data, potentially affecting the accuracy of our measurements regarding which areas of the screen captured the most attention. These factors underscore the importance of exercising caution when generalizing our results to broader gaming contexts. To enhance the reliability and relevance of future research, we recommend that future studies should include controls for these individual differences and consider a wider variety of game types. This approach would not only offer a more comprehensive understanding of how advergames impact brand recall across different gaming environments but also help refine the methodologies used in eye-tracking studies to accommodate variability in participant ergonomics.

5. Conclusions

The final observation was that players generally focused on the screen at their eye level during advergame play. Consequently, it was advisable to position advertised brands or products at or slightly above the middle level within the game content to enhance visibility compared to other positions.

These findings offered several insights for future research in this domain. The study suggested that advergames were versatile tools for advertising any brand, product, or company message. However, the dimensions of Action–Awareness Merging and the Transformation of Time were found to be unreliable for both the experimental and control groups. It would be beneficial to reevaluate and modify the questions related to these dimensions for future computer-based games. To more accurately analyze the Transformation of Time dimension, researchers could measure players’ perceptions of time by asking, “How long did you play the game?” and then compare their responses to the actual gameplay duration. The study also revealed that positioning products on the middle and top shelves increased their visibility to players, a strategy that could be applied in real-world settings. Additionally, the relationship between eye-tracking fixation duration and brand recall could be explored in 2D games, where aggregate eye movement data could be more easily collected than in 3D games. The aggregated eye movement data in this study was found to be sufficient to establish a relationship with the number of brands recalled.

In our study, we intentionally chose simple games to eliminate any additional variables and concentrate on the specific impacts of brand recall and attention. However, we do recognize that this simplicity may have affected the level of immersion for the participants, which is a vital aspect in fully comprehending the influence of advergames on user engagement and brand recall. Simple games may not completely capture the deep engagement that more complex and visually stimulating games can provide, potentially limiting the general applicability of our findings to the intricate gaming environments typical of the current market. This limitation implies that while our results offer valuable insights, they may not fully represent the dynamics of player immersion in more sophisticated gaming contexts. Future research should contemplate utilizing more complex and commercially viable games that more accurately replicate the immersive experiences prevalent in modern gaming. Such studies would not only validate our findings but also broaden their scope to encompass scenarios that are more consistent with the experiences of contemporary gamers, thereby enhancing the applicability and significance of the research in real-world settings.