I
Programmatic
Advertising’s Effect
on Consumer
Decision
Artificial Intelligence Technologies in Advertising and
Marketing on Consumers Decision Making
BACHELOR’S THESIS WITHIN: Business Administration NUMBER OF CREDITS: 15
PROGRAMME OF STUDY: International Management AUTHOR: Hilal Ahmad & Sepehr Teimouri Mokarram
Acknowledgments
We would like to thank our tutor, Gershon Kumeto for all the consecutive efforts and commitment he implemented to help us with the thesis and the great feedback given by him. With his help, we gained a lot of insight and motivation towards achieving our goal.
Furthermore, we would like to thank our fellow students who gave us critical and beneficial feedback which allowed us to gain an insight into the path of our paper and the goal that we were going towards.
Last but not least, we would like to thank our friends and colleagues who also provided us with feedback as well as students within our seminar group.
______________ ______________
Bachelor Thesis in Business Administration
Title: Programmatic Advertising Effect on Consumer Decision Making: Artificial
Intelligence technologies in Advertising and Marketing and its Effect on Consumer Decision Making
Authors: Hilal Ahmad & Sepehr Teimouri Mokarram Tutor: Gershon Kumeto
Date: 2015-05-22
Key terms: Consumer Decision Making, Programmatic Advertising, Tailored Advertisements
Abstract
Background: With the ever-growing advancement of technology and the implementation of new groundbreaking technology in our day-to-day lives, a path of curiosity was opened that attracted attention. Just how much does the use of artificial intelligence (AI) affect consumer behavior and how much do consumers trust AI.
Purpose: The purpose of this thesis is to explore the effects of the implementation of AI and machine learning in marketing on consumer behavior and measure the level of trust
consumers have towards this advancement of technology.
Method: This research study is conducted through a qualitative method while taking advantage of interviews from individuals based in Sweden ranging from ages of 18 to 30 carried out in a thematic analysis approach.
Conclusion: the results show that implementation of AI in marketing has direct effect on the consumer behaviour. The authors have used various different of primary data collected from a number of interviews and secondary data from depth research of articles as well application of consumer decision Process (CDP) to analyse and evaluate the findings.
Table of Contents 1. INTRODUCTION ... 1 1.1. BACKGROUND ... 1 1.2. PROBLEM DISCUSSION ... 2 1.3. RESEARCH PURPOSE ... 4 1.4. DELIMITATION ... 4 2. FRAME OF REFERENCE ... 5
2.1. AIPROGRESS OVER THE YEARS ... 5
2.1.1. ETHICS ... 6
2.2. THE ADVERTISING INDUSTRY ... 6
2.2.1. ROLE OF ADVERTISING ... 6
2.2.2. THE ONLINE ADVERTISING INDUSTRY ... 7
2.3. PROGRAMMATIC ADVERTISING (INTELLIGENT ADVERTISING) ... 7
2.3.1. SOCIAL MEDIA ... 8
2.4. AI-HUMAN INTERACTION ... 8
2.5. DILEMMAS WITH AI-CREATED CONTENT ... 9
2.6. BUILDING TRUST AMONG CONSUMERS ... 10
2.7. CONSUMERS’WILLINGNESS TO ACCEPT AI-CREATED ADVERTISEMENTS ... 11
2.8. CONSUMER DECISION PROCESS ... 11
3. METHODOLOGY ... 13 3.1. RESEARCH PHILOSOPHY ... 13 3.1.1. RESEARCH APPROACH ... 13 3.1.2. RESEARCH DESIGN ... 14 3.2. METHODS ... 14 3.2.1. DATA COLLECTION ... 15 3.2.2. PRIMARY DATA ... 15
3.2.3. SEMI STRUCTURED INTERVIEWS ... 16
3.2.4. INTERVIEW QUESTIONS. ... 16
3.2.5. DATA ANALYSIS ... 17
3.3. ETHICS ... 17
3.3.1. ANONYMITY AND CONFIDENTIALITY ... 17
3.3.2. CREDIBILITY ... 18
3.3.3. TRANSFERABILITY ... 18
3.3.4. DEPENDABILITY ... 19
4. EMPIRICAL FINDINGS: ... 20
4.1. CHANGES IN NEED ... 22
4.1.1. PERSONALIZED ADVERTISEMENTS ... 22
4.1.2. CHANGE IN MODE OF ADVERTISING ... 23
4.2. CHANGES IN REQUIREMENTS ... 24 4.2.1. PERSONALIZED ADVERTISEMENTS ... 24 4.2.2. TIME EFFICIENCY ... 26 4.2.3. BASIC UNDERSTANDING OF AI ... 26 4.3. SHIFT IN BEHAVIOUR ... 27 4.3.1. PERSONALIZED ADVERTISEMENT ... 27
4.3.2. CHANGES IN PURCHASE BEHAVIOUR ... 28
4.3.3. ATTITUDE TOWARDS AI ... 28 5. ANALYSIS ... 29 6. CONCLUSION ... 33 7. DISCUSSION ... 35 7.1. CONTRIBUTIONS ... 35 7.2. PRACTICAL IMPLICATIONS ... 35 7.3. LIMITATIONS ... 35 7.4. FUTURE RESEARCH ... 36 REFERENCE ... 37 APPENDICES ... 43
1. Introduction
1.1. Background
Media in this day and age is completely digital, with everything from newspaper to television going online on some sort of digital platform. Not only that but the non-media experience is also reliant on digital devices, for example, communication with peers through cell phones and computer-mediated communication through instant messaging services and emails (Sundar, 2008). Hence it wouldn’t be wrong to assume that the marketing and advertising industry does not take advantage of this technological and digital transformation.
Advertising is made and designed in a certain way to promote the sale of a product or service. It has however been around for a very long time, dating back to ancient times in many
different cultures and civilizations. This compared to today’s more used form of advertising also known as programmatic advertising is defined by Deepa Naik (2019) as “Programmatic Advertising is the automated buying and selling of digital advertising through centralized computer-driven ad exchanges and related databases and management platforms.” While speaking on the subject of Advertising and its transformation over the years Deepa Naik (2019) explains that earlier over the internet buying the advertising space was simple, where there were only two parties, the website, and the advertisers, for the process of buying and selling. Paid advertising later had several other options such as Sponsored content, Google AdWords, Bing, etc., where the marketers had to make a judgment on how to allocate their advertising budget. But this all changed with the advertising industry becoming more
advance with the implementation of automation and machine learning (machine intelligence) and gave birth to Programmatic Advertising.
Machine learning can also be seen as one of the branches of Artificial Intelligence (AI). Machine learning is given a task based on the data it has. After the task has been performed it waits for feedback, if the feedback is negative it then requires more information to correct its wrongs and if the feedback is positive it tries to duplicate it (Sterne, 2017). According to Davenport (2020), AI is the activity and process conducted to make
machinery intelligent, and intelligence is the attribute that allows that system to function with foresight in its environment. To understand how AI and machine learning work, they explain that AI is used in a vast majority of tasks, for instance, game playing. There are machines that can play master-level chess, the AI planted in these machines works through algorithms, looking at numerous positions and calculating the best possible move. AI with the help of machine learning mimics human intelligence. The integrated computer with these systems will be able to learn how to respond to different types of actions therefore it uses numerous algorithms and data to form different predictions. Moreover, AI is capable of doing so much more than just forming algorithms to predict different data, but this is the most commonly used functionality of AI for marketing and business.
Marketers take advantage of so many features of AI. One of the most used tactics by marketers is using search engines. For instance, when someone searches for burgers online, AI has to decide if that person is looking for information regarding burgers or they are looking for the best and nearest hamburger place. Furthermore, the ads that are circulating on every online platform are shown to different people based on their personality and the AI system decides which ad is for which sort of audience. Another example of how AI is engaged with ads and individuals’ daily life, is when a certain product or service is being searched online. AI sees the algorithm and shows ads about that certain product or service to consumers on different platforms, such as Instagram and Facebook, that are heavily
influenced by advertisements. (Harvard business review, 2016). Cambria et al., (2012) state that machine learning that is used in specialized AI software which are based on algorithms enable to determine specific patterns within big data and to categorize them, this is specially used in the deep analysis of social media content.
1.2. Problem Discussion
It is understandable that shifting to a new age of digital transformation is coming with
concerns and is making a population of people worried about the future (AI magazine, 2012). However, AI and machine learning could be seen by many as a great advancement and milestone of technological use in the field of marketing and advertisements. Despite the advancements and achievements there still are some significant concerns that are yet to be addressed. These concerns could be of privacy and transparency of the data being collected,
the accuracy of the targeted marketing and advertisements towards the consumers, and its effect on the consumer behavior.
Li (2019) stresses the matter of data privacy and the importance of this issue, they mention that the data set include consumer’s personal identifier, biometric data, geolocation, internet browsing history, psychometric data, and the deductions that a company might have made on a consumer. These are all key information that is vital for AI applications in the process of advertising, this is especially crucial for generating brand messages and discovering
consumer insights. They further mention that it should be of the utmost importance that the consumer knows what personal information is being collected and by which entity. The consumer should have the ability to control their data and be able to choose if they want to authorize access to someone in return for some benefit. Trust should be established between both the entity and the consumers since these intelligent ads are made to serve users in a better way.
According to Bird et al., (2020) “AI is created by humans, which means it can be susceptible to bias. Systematic bias may arise as a result of the data used to train systems, or as a result of values held by system developers and users.” This can be stated as a usual occurrence
because the machine learning applications are often trained on the data sets that are reflective of a specified or certain demographic group and or of that is reflective of societal bias. They also mention that several cases have been brought to attention where unintended social bias has been promoted, that has been automatically reinforced or reproduced by AI systems. Zhau et al (2017) shed light on this by giving an example of when women searched for wedding dresses, they were shown traditional western white dresses whereas the Indian wedding dresses were categorized as “performance art” dresses.
Another negative stereotype of machine learning could be its lack of experiencing emotions and sensations that differentiate machines from human beings (Haslam et al., 2005). “This negative stereotype tends to trigger feelings of discomfort or eeriness when the boundaries between humans and machines become obscure, such as when a machine behaves like a person or does a human job” (Wu and Wen 2021). Wu and wen (2021) also mention that people have some idea that the advertisements messages are created by AI and tend to be more accurate. However, they still find this unsettling as it is perceived by them that the
1.3. Research Purpose
Most people have access to internet through their mobile devices or computer, either for work or pleasure. Because of this, they are usually exposed to some sort of marketing tactics. As mentioned above most marketers nowadays use AI systems to target their potential audience. Therefore, the purpose of this study is to further expand the research on the influence of AI in the context of marketing and advertising on consumers. Hence, the research question for this study being:
“How does AI in marketing and advertising influence consumer’s decision behavior?”
1.4. Delimitation
This paper will focus on consumers in the Swedish market, in terms of their knowledge, awareness, and concerns on the topic. To start, the research is only limited to consumers in Sweden, as the respondents are easily accessible here. Secondly, the research carried out in this paper will be limited to investigating the influence of marketing and advertising through AI systems has on consumer behavior.
2. Frame of Reference
Literature Review
Databases Jönköping University Library Services Primo; Google Scholar;
Main Theoretical Fields Consumer Behavior, Artificial Intelligence in Marketing and Advertising, Digital Transformation
Search Words Programmatic Advertising, Artificial Intelligence; Marketing, the influence of Artificial Intelligence on consumers. Consumer behavior, Digital Transformation Literature Review Academical Articles, Journals and Literature books, and
the internet Language of publication English
2.1. AI Progress Over the Years
Rosenfeld and Richardson (2019), states that as AI has matured and became ubiquitous, there has been a developing evolution of systems where people and systems work alongside each other. They further elaborate that these systems are come to be known as human-agent systems or human-agent Cooperatives, and they have moved from theory to reality in the forms of digital personal assistants, recommendation systems, service robots, self-driving cars, chatbots, planning systems, and training and tutoring systems. Azaria et al. (2014), explain that the automated agents are designed to strategically disclose pieces of information that encourage humans to take some actions over others. Roesenfeld and Richardson (2019) raise a concern about the explainability of the human-agent systems and sets three criteria on it based on its usefulness, benefits, and criticalities. They explain one of the well-rooted concepts within human-agent or human-robot groups that is adjustable autonomy. This refers to the amount of control that an agent or robot has in comparison to human users. If the agent is solely controlled by a human operator, then no explainability is needed. However, if full control is given to the robot, specifically if the reason for this choice is obvious. For example, recommendation systems/agents that give suggestions and advice based on a well-defined collaborative filtering algorithm, again no explanation is required. In contrast to this, they provide other human-agent systems that are built through which agents' function is to assist a human task. In these cases, an agent’s explanation is thus in most cases a critical element
within the system. Hence, task execution is directly linked with the need for the agent to be completely transparent or it should be able to faithfully and explicitly explain its action. They suggest that an active explanation might act as a critical factor needed to make a life-or-death decision within human-agent systems.
2.1.1. Ethics
Hagendorff (2020) mentions that over the years a great deal of ethics guidelines has been issued in contrast to the advances in research, application, and development that are being done in the context of AI. These guidelines are aimed towards the normative principles and recommendations to tackle the disruptive power of the new AI technologies. They further on explain that AI ethics or generally ethics lack the mechanisms to stress its normative
challenges. The implementation of ethics is often associated with reputational losses in terms of misconduct, or constraints on memberships in specific professional groups. Generally, they do not pose any prominent threat because these mechanisms are rather weak. Therefore, researchers, consultants, activists, managers, and politicians have to deal with ethical weaknesses. Yet, this results in making ethics more interesting for AI institutes and companies, as they have to constantly formulate their ethical guidelines. Companies and institutes when integrating ethical considerations regularly in the areas of public relations or implementation of ethically driven self-guarantees are continuously being discouraged to form a legally binding framework. Hagendorff (2020) portrays that the abundance of specific laws to diminish the possibilities of technological risks and to eradicate the scenarios of misconduct is justified because of the self-governed ethics guideline of the AI industry. However, it is deemed necessary to demand more concrete laws for the AI systems as these demands yet remain superficial and vague. This raises the question of trust that the
consumers have in the services and platforms that are using AI technologies.
2.2.The Advertising Industry 2.2.1. Role of Advertising
The main purpose of advertising is ultimately designed in a certain way to improve and enhance the sales of different goods and services in different ways. Some advertising is generated in a way to increase the sale numbers by identifying leads (Qin and Jiang, 2019). Other advertising is informative. It provides the consumers with information about prices and products, which they can eventually use to make the purchasing decision. Li (2019) As an
example of informative advertising, are newspaper ads that list supermarket items alongside their listed prices.
2.2.2. The Online Advertising Industry
The online advertising industry is based on selling and buying advertising space that is accessed by viewers through the internet. There are 4 main sections of online advertising divided by the industry observers. These four sections are as listed:
1: search advertising: this the type of advertising that appears on search results pages. 2: display advertising: appears on non-search web pages.
3: classified listing: that appears on websites. 4: internet e-mail-based advertising.
In many ways, online advertisement has a lot in common with traditional ways of advertising. Publishers take advantage of content to enchant viewers and then sell advertisers access to those people viewing them. Advertisers are enabled to place text, pictures, or video ads in the spaces provided by the suppliers.
In many ways, online advertising is very similar to traditional advertising where publishers can use different content to attract consumers to products or services and then sell the rights and access to those viewers.
2.3. Programmatic Advertising (Intelligent Advertising)
Xu et al. (2020) explain AI in the context of consumer Service-specific as "a technology-enabled system for evaluating real-time service scenarios using data collected from digital and/or physical sources to provide personalized recommendations, alternatives, and solutions to consumers’ inquiries or problems, even very complex ones." Bakpayev et al. (2020) as well sheds a light on how continuous programmed and interactive advertising has evolved digital brand communication. Shankar et al. (2010) further expand on this by stating that the face of the communication environment has evolved from marketers-to-consumer (one-way) to a more interactive one, where two or multiple ways of communication are used and seeking active consumer engagement through automated conversations. They further suggest that the communication industry has been able to reshape itself because of AI. With the help
of AI, it is now possible to penetrate various facets of the advertising process. These may include market analysis and research, design and copyrights, performance assessments and feedback, as well as media planning and buying (Qin and Jiang, 2019)
Li (2019) suggests that intelligent advertisements not only illustrate communication at an abstract level but also portray its possible potential for the future of advertising. They refer to this as “Consumer-centered, Data-driven, and algorithm-mediated brand communication.” Taulli (2020) Explains that programmatic advertising has made it possible to be able to retain efficiency in terms of message creation and delivery. Because of this intelligent advertising conceiving more exposed and personalized brand communication through the collection of consumer data has become easier. Jenning (2019) sheds light on this phenomenon by using an example of TikTok, a video-sharing social media platform, which already has a consumer base of over 800 million around the globe and still growing. Tiktok uses Machine learning and AI to learn about their consumers’ preferences and interests through their interaction with the application.
2.3.1. Social Media
Sterne (2017), while talking about the use of AI in marking explains that marketers use AI technologies to monitor social media effectively. This gives them a clearer picture of what is being discussed by the public about a brand through their posts and comments, this is known as sentiment analysis. They also help with, audience analysis, determine how to reach them with personalized content. In addition to this, image analysis also proofs to be useful for marketers to analyze brands that are active on social media based on their recognition of the images of the logos being shared. Ashley and Tuten (2015) further elaborate that marketers are fully optimizing sentimental, image, and audience analyses through AI tools. This helps them to identify which branded content received the most amount of consumer engagement through social media.
2.4. Ai-Human Interaction
Sundar (2020) explains AI-Human interaction as where people adapt to media sources with the perception of an intelligent entity is efficient enough to modify content in unprecedented ways. According to Sundar (2008), the modality agency interactive navigability (MAIN) model consumers' evaluation of media content can be influenced and could be triggered by
some cues that are presented on the media interface. Fiske and Taylor (1991) explain that this is because people tend to think cognitively and opt for effortless options to solve problems and think instead of a more sophisticated and more effortful one. Wu and Wen (2021) state that when people find an easier way to take mental shortcuts to process information, they tend to stray away from effortful ways of analyzing information.
2.5. Dilemmas with AI-Created Content
Advertising has become more sophisticated and digitalized over the years, where
communicating the right message or information to the right person at the right time has become crucially important (Adam, 2004). AI has revamped the understanding of consumers’ needs, wants, and desires for organizations and marketers. Now businesses and marketers can refine and predict consumers’ considerations sets that are based on consumer interaction through leveraging data-derived consumer preferences (Esch et al. 2021). They further expand on this by stating that businesses are compelled to re-examine and evaluate their existing practices in this new phase of digital messaging and invest in AI. Kietzman et al. (2018) explain this through a UK-based fashion e-commerce business called ASOS, that uses AI combined with machine learning, along with natural language and image generation in order to learn the browsing activities of the consumers. This learning mechanism facilitates refined and tailored advertising while increasing its effectiveness.
According to Ene (2018), the use of AI has become one of the most important elements to be used by businesses and managers in order to keep up with the competition and be successful. However, with the implementation of AI, the changes in the internal systems and
management styles are not always smooth and natural, and often generate serious problems and issues for the company and its interaction with the consumers. They also stress the fact that the implication of technologies as such in various forms of business has raised an ongoing discussion on the willingness of people to accept these technological trends. The development of these technological advances has many benefits for the company in terms of cutting costs and being more efficient as compared to using a human employee. With this, they are able to achieve a competitive advantage in terms of being available to consumers 24/7 through the help of innovative and automated systems (Ene, 2018)
However, the challenge with this according to East et al., (2008) is to increase consumer satisfaction, as compared to humans, robots do not have all the answers and are not as flexible while interacting as humans are. Also, in case of a system failure they are not available, hence explaining this to an impatient customer waiting on call would be difficult. Lee and See, (2004) analyze the impact of these technologies and people’s willingness to accept and use these technologies. They explain that if initially, a system is untrustworthy, it will not be used. Hence it will not reach its full potential if it’s not being used and because of this, the confidence in the system is quite unlikely to grow. McKnight et al., (2002)
contributed to this phenomenon by stressing the initial confidence in these technologies in the initial phase of implantation which portrays its reliability and over time shifts it to belief. Eventually leading the subject to trust and accept it in the final phase. Moreover, when a consumer behavior switches due to trends, it is rather difficult to enhance consumer behavior through automatization (Jia, 2009)
2.6. Building Trust Among Consumers
It is very difficult to define trust. Trust has a trumping effect. When the term trust is mentioned in a dialogue, it has a demoting effect on the status of other values.
In other words, trust is an emotional state for the brain so it could not be explained through just expectation of behavior (Baier, 1994).
Trust is known to be the core part of all human relationships, which includes romantic partnerships, family life, business operations, politics, and medical practices. If you aren’t able to trust someone or something, it makes it difficult for you to be able to benefit from their advice whether it is professional or not. Trust is a package of behavioral threads which includes acting in ways that depend on another. It’s the belief in the probability that a person will behave in certain ways (Acton, 1974).
One of the biggest assets of a business is the trust their consumers have towards their business. Without them, no business would thrive and succeed (Smith, 1790). Ignoring consumers and seeing them as just numbers rather than actual humans can be detrimental to the successes of any business. There are plenty of reasons why building and maintaining trust between consumers and organizations is vital.
Any organization without the demand of their consumers is worthless and would simply fail to flourish regardless of the product or service they provide. In general, it’s the consumers who control the success of companies. If a company is not credible and is hard to trust, it is most likely that consumers will avoid that organization altogether.
There are many ways for marketers and advertisers to build trustworthy relationships with their consumers. Establishing trust could be possible through factors like:
making sure the business is as secure as possible, signing up to as many relevant social media accounts as possible (Facebook, Instagram, Twitter, etc..), and having a professional user-friendly website.
2.7. Consumers’ Willingness to Accept Ai-Created Advertisements
Wu and Wen (2021), in their paper mention that AI technology has been utilized for the creation of advertising messages. They also state that several factors influence consumers to appreciate AI-created advertisements. Not only that but consumer perceived objectivity towards the AI-created advertisements process was influenced positively towards the
machine heuristics, that machines are more trustworthy than humans. A boost in consumers’ appreciation towards AI-created advertisements is also a result of this. Their findings
indicated that “consumers’ overall perceptions of advertisement creation as an objective process and feelings of unease with robots are important antecedents that determine their appreciation of AI-created advertisements through the psychological mechanisms of machine heuristic and perceived eeriness. These findings are believed to provide both theoretical and practical insights that help deepen understanding of AI advertising.”
2.8. Consumer Decision Process
A Decision-making model is usually very complex and requires many steps. However, this model may be the most ration and for sure the more complex than the purchase situation warrants (Sethna and Blyth, 2019). This model further developed and became CDP (Consumer Decision Process). CDP consists of the following steps
1. Need Recognition: The individual recognizes that something is missing from their life.
2. Search for information: This information search can be internal (remembering facts about products or recalling experiences with them as a result of services) or external surfing on the internet, reading about possible products, or visiting a shop)
3. Pre-purchase evaluation of alternatives: The individual considers which of the possible alternatives might be best for fulfilling the need
4. Purchase: The act of making the final selection and paying for it. 5. Consumption: Using a product to fulfill a need
6. Post-consumption evaluation: considering whether the product satisfied the need or not and whether any problems were arising from its purchase and consumption 7. Divestment: Disposing of the product, or its packaging, or any residue left from
3. Methodology
3.1. Research Philosophy
The research conducted for this paper is an exploratory qualitative study. To maintain the authenticity of this paper only peered review articles are reviewed that are accessed from Jönköping University’s online library (primo) and google scholar. Another supporting material is also used from well-recognized news platforms and websites, such as Forbes, times, world economic forum, etc. In addition, interviews are also held to gain first-hand insights into the perception of AI. This along with the literature review provides a cohesive understanding of the factors involved in the context of trust in AI from the consumer's point of view.
Methodology within research is explained as the systematic method to resolve a research problem by gathering data while using various techniques while providing an interpretation of data gathered and drawing conclusions about the research data. Fundamentally, a research methodology is the blueprint of research or study (Murthy & Bhojanna, 2009).
3.1.1. Research Approach
Commonly, an interpretive approach is known as the understanding that has been based upon feelings and believes from individuals who are participating in the study. It is also known as interpretivism involves researchers interpreting elements of the research, therefore
interpretivism demonstrates human interest in a study (Preston, 2006).
Accordingly, interpretive research is believed to access reality whether it is given or socially constructed through socially constructed methods such as language, consciousness, shared meaning, and instruments (Collins, H. 2010). Development within interpretivism philosophy lays within the critique of positivism in social sciences (Collin, H. 2010). Appropriately, this philosophy emphasizes mostly qualitative analysis rather than quantitative analysis.
Interpretivism is correlated with the position of idealism within the philosophy and is used to group diverse approaches like, social constructivism, and phenomenology approaches that reject the objectivist view that resides within the world of consciousness (Littlejohn, S.W. & Foss, K.A. 2009).
According to the interpretivism approach, it is vital that the researchers as social actors appreciate differences between individuals (Saunders, et al., 2012). Furthermore,
interpretivism studies mostly focus on meaning and engage multiple methods in order to be able to reflect different sides and aspects of the issue on hand.
Fundamentally, the interpretivism approach is based on a naturalistic approach to data collection. For example, interviews and observations. Another type of data research that is also popular with interpretivism philosophy is secondary data (Collins, H. 2010). In this sort of study, meanings emerge mostly towards the end of the research process.
The main advantages which are regarded with interpretivism relate to the subjective nature of this approach and the potential for bias on behalf of researchers. The primary data generated in the interpretive studies are not available to be generalized since data is heavily influenced by the personal point of view (Myers, M.D. (2008).
On a positive note, however, a beneficial influence of interpretivism on qualitative research is on the cross-cultural differences in organizations and issues of ethics. Interpretivism helps study these factors in greater depth. Therefore, the primary data that has been collected to study through interpretivism comes with a high level of validity because data in such studies tend to be very honest and trustworthy.
3.1.2. Research Design
To be able to answer the research question sufficiently, it is crucial that the structure and design of the research is established with its purpose in mind. Therefore, qualitative research approach is considered to understand and examine the data collected through interviews efficiently as well as the theories and literature review.
3.2.Methods
In this research, the authors have decided upon a methodical review to examine the effects of AI within the advertising industry and consumer trust. Both primary and secondary data that has been collected during this research would be combined to generate honest research which
is easily comprehended. The reason for not opting for a quantitative approach was because of its complexity with the subject material in context with the subjectivity and perspective nature. Also, correlation testing would also prove difficult due to complex question and making their estimations. Researchers have more freedom and flexibility while using a qualitative approach. Giving the researcher the ability of interpretation of the data and in-depth understanding of the consumers behavior. A thematic approach is chosen to analyze these findings.
3.2.1. Data collection
Data collected throughout this study is formed through two types of data which are the primary data and secondary data. Primary data are data gathered for the research question at hand, using methods that fit the research in the best possible way (Thompson, 2000). Every time primary data is being collected, fresh data are added to the existing store of social knowledge.
Primary data is mostly collected through interviews, questioners, surveys, and direct observations.
Secondary data, on the other hand, is the primary data that has been previously collected by someone else and is being put to use. The use of secondary data is usually easier and less time-consuming that is because it has already been collected and hence easily accessible (Silverman, 2000).
The most crucial difference between these two types of data is that primary data has been carefully collected regarding a research purpose and serves exactly towards that purpose therefore secondary data seldom could be used as a source of primary data. With that in mind, neither one of the mentioned methods of data collection could be announced better since both forms of data are crucial in research and provide separate beneficial information. Combining these different forms of data will help to provide research that is both honest and rich in information (Thompson, 2000).
3.2.2. Primary data
transcribed by the authors before applying it in thematic structure which will increase the credibility of the primary data collected. The authors have collected the data through an interpretive paradigm. To make sure the interview is carried out in an honest matter and the interviewees have given fair answers, the participant in the interview are asked to keep their identities completely anonymous.
3.2.3. Semi structured interviews
While the study was being carried out, a total of 10 interviews were done by mid-April. According to the research, the authors decided on a semi-structured interview in which the participant to the interview was handed a sheet containing the interview questions while follow-up questions were being asked if there were any need for them. This style and format of interview allowed the authors to ensure that the answers were given accurately and straightforwardly to each question hence increasing the credibility of the interview. Before the start of the interview, the authors first introduced themselves, and then the
participants of the interview were presented with a consent form for them to be signed before proceeding.
The duration of the interviews was between 25 to 40 minutes and due to the pandemic of the coronavirus, it was decided that they be carried out through zoom calls and phone.
Participants were in age groups of between 18 to 30 years. The language used to conduct the interviews was English.
3.2.4. Interview questions.
A full list of the questions of the interview will be posted and available in the research. The participants of the interview were asked a series of detailed questions about the effects of AI on marketing and the level of trust consumers have towards it. The questions were designed in a way to gain useful answers from the interviewees. A topic would first be discussed by the authors, the interviewees would then answer in a general manner, and then follow-up short questions would be added for the candidates to elaborate more on the matter and address the interview while at the same time avoiding unnecessary specific answers.
Consumer decision behavior was utilized as the theoretical framework in this research to understand how general day-to-day consumers feel about the application of AI within the advertising and marketing field. To achieve and maximize this the interview questions were
formed around the mentioned theoretical framework in a way to determine and measure the level of trust consumers have towards AI in marketing and advertising.
The theoretical framework of S.W.O.T was considered for the research but after careful consideration and investigation, it was decided upon by the authors that it is not a feasible option since it would not provide any beneficial use towards the topic of the research. It was concluded by the authors that the S.W.O.T framework would investigate the strengths, weaknesses, and other aspects of AI and would push the research to be rather technical where else the main purpose of the research was going to be about the consumer decision behavior.
3.2.5. Data analysis
The purpose of this research is set to revolve around a qualitative data analysis therefore a thematic analysis method has been selected to evaluate the collected data.
With the selected method of data analysis, the reader will be provided with a more vast understanding of the effects that AI has on marketing and advertising and the significance of the trust consumers have towards Machin learning and AI.
Due to the thematic analysis, the authors have carefully inspected the interview data to produce the most comprehensive and accurate data and meaning.
3.3. Ethics
The quality of the study is correlated with maintaining and considering a higher standard of ethical approach while conducting the research. The authors considered all the necessary steps to be taken to ensure the credibility and reliability of this report. Anonymity and confidentiality, credibility, transferability, dependability, and conformability all these factors were considered in this report
3.3.1. Anonymity and Confidentiality
Participants that took part in this study and answers were promised to be kept anonymous for their safety, through this it was possible to achieve the most information and feedback from the participants. Saunders et al., (2012) address that a researcher should be able to assure its participants about the safekeeping of their identity context to the information that they
provide. Therefore, establishing a trust between the participants that the information provided by them would not lead back to them, thus providing them with the protection of anonymity
and rights. Participants before conducting interviews were made aware of the purpose of the study and were made aware of their rights in regards to the privacy of their data being collected. All the subjects in this study participated voluntarily and were given a consent form to sign after they were informed about their rights to withdraw, share or keep the information they provide during the interview to the authors. All participants were asked for their permission to record the interview, to which all 10 of the participants consented. They were also informed that these recordings will later be deleted upon the submission of the thesis. To keep the anonymity of the participants they were informed about the allocation of numbers from 1 to 10 to them, relation to the order the interviews were conducted at.
3.3.2. Credibility
According to Shenton (2004) and Bryman & Bell (2007), credibility is associated with trustworthiness and describes the intention of the conclusions drawn while examining a phenomenon or drawing a conclusion concerning the reality in a paper. Saunders et al. (2012) also mention that if the key information regarding the study is conveyed to the participants it increases the credibility of the research. All the participants were beforehand informed about the topic of the thesis along with its key themes and other necessary information. In addition, to ensure the credibility of the paper, it should be able to provide similar results if the same method is applied. The participants were provided with the opportunity to be able to provide answers and arguments from their own experience due to semi-structured interviews. To avoid any mistakes while coding and analyzing the results, the interviews were recorded and transcribed. The semi-structured interviews enabled the participants to express their
arguments, rationalizations, and thoughts, therefore it aided in viewing specific topics from different angles and more in-depth. However, the semi-structured interviews did cause a bit of instability and hindered the credibility of the findings, as some deviation could be
observed between responses to some extent.
3.3.3. Transferability
Sunders et al. (2012) state that transferability refers to the extent to which findings acquired from a qualitative study can be applied to other studies within a similar context or to a level where they can be generalized. As this study focuses on a smaller segment of individuals who are Swedish citizens, therefore the transferability and generalizing of this study could be limiting. In addition to that this study might not be of relevance to the other part of the world
with different cultures with its findings and results. However, a similar study with a larger sample size could use the given frame of reference with the provided results in this thesis.
3.3.4. Dependability
Cope (2014) states that dependability is the consistency of the data in similar conditions. If the researcher’s process and description are used in a similar setting, it is deemed dependable if the study’s results are replicated. Bitsh (2005) further elaborates that if all the research process is clearly and well documented along with the decisions taken in terms of
methodology, researchers can maintain the dependability of their findings in the context of a qualitative study. For this study, all the interviews were recorded and then transcribed as mentioned previously. After transcribing they were matched with the recordings of the participants to ensure none of the information is being left out. These transcribe were later used to code and analyze using thematic analysis. Later all the data was analyzed by all the authors to search for any inconsistencies, hence improving the dependability of the research. The authors to their best of capabilities avoided any biases that might occur from comparing the empirical findings with the frame of reference and did not aim for any fixed results for this study.
3.3.5. Confirmability
Confirmability refers to the ability of the researcher to demonstrate that the data represents the responses of the participants and not the viewpoints or biases of the researcher (Cope, 2014). Therefore, the authors of this report confirm that all the data collected for this study were retrieved in the form of interviews and literature review. To increase the credibility of this report only peered reviewed articles along with other academic literature in the form of books were used. All the assumptions were avoided to maintain the integrity of this study. However, since the nature of this report being exploratory few assumptions were made in the conclusion. To increase the credibility of this report quotes from the interview are also used in the study.
4.
Empirical Findings
:This section will present the empirical findings that have been obtained through
semi-structured interviews. The following themes were identified through thematic analysis of the interview data. Change in need, change in requirements, and shift in behavior. Through these authors were able to determine six sub-themes from the above-mentioned main themes. Along with these three main themes the authors also found six more underlying sub-themes where personalized advertisements were found to be common in all of these main themes. To further explain how these themes and sub-themes were developed through the coding
process, a dissect is provided under table 1. To make it easier for readers to understand and read, hey would be distributed in different main themes and sub-themes for every category.
This section will also present all the participants and their answers. Table 2 below presents all the interviewees that our study has focused on. The participants are numbered from 1 to 10. In addition to that, the duration, date, and mode of interviews are also presented in table 2 below.
Table 1: Extract of thematic analysis
Main theme Sub-themes Quotes
Changes in Need Change in mode of Advertising
“We need to adapt to the new way of advertising because we don’t watch TV anymore, we’re totally different consumers now than we were 10 or 5 years ago.”
Personalized Advertising
“…I don’t have to look for things that I need like I’m working and have kids, so the system knows what I’m interested in and I only get those ads and do not get very general ads if you know what I mean, which is very nice.”
Change in Requirements
Personalized Advertisements
“I like it, sometimes you get new ideas, like things that you didn’t know existed.”
Time efficiency “In my opinion, an ad should be able to deliver its message within the first 5 seconds, that’s when you grab the most attention, in my opinion”
Basic
understanding of AI
“I know that AI is being used in marketing today, especially with algorithms made by big companies such as Facebook”
Shift in Behavior Personalized Advertisements
“Sometimes I would look for a product but have no luck finding it and later on I would come across it on a social media ad and purchase it”.
Changes in
Purchase Decision
“at times I would be looking for a certain type of product online and then see an ad for something that is reletively different and purchase that instead”
Attitude towards AI
“it would be a lot better if it was made clear to users which data they were collecting from
them and how were they collecting these data”
Table 2: Participants interviews
Participants Date of Interview Duration Interview Type
1 27/04/2021 33 min In-Person 2 27/04/2021 29 min Zoom 3 28/04/2021 39 min In-Person 4 29/04/2021 28 min Zoom 5 29/04/2021 32 min Zoom 6 30/04/2021 37 min In-person 7 02/05/2021 29 min In-person 8 02/05/2021 31 min Zoom 9 02/05/2021 27 min Zoom 10 04/05/2021 36 min Zoom 4.1. Changes in need
This first main theme that is identified by the authors through thematic analysis is changes in need. The authors dissected this drastic change due to the evolution of technology and how consumers like to consume information that is more in line with the technological
advancements. The below define categories that were found through the process of coding of the thematic analysis will further elaborate on it.
4.1.1. Personalized Advertisements
This sub-theme identified was the most reoccurring one that the participants identified during the interviews. Most of the participants acknowledged that they have seen a shift in their need based on the content of the advertisements that they come across.
Participant 3 state that: “Nowadays, especially during this time of the pandemic, I have
reverted to only shopping online, and I feel like these targeted advertisements is something that I appreciate. Like I don’t have to look for things that I need like I’m working and have
kids, so the system knows what I’m interested in and I only get those ads and do not get very general ads if you know what I mean, which is very nice.”
Participant 7 adds to this by stating: “There is so much stuff on the internet so obviously I know I cannot know about everything, but I don't feel like I'm missing out, since I get very personalized ads that are of my interest.”
These participants shed light on how there was a much-needed change in the advertising industry as they want to see more and more tailored and personalized advertisements that are of their interest and not something very generic.
Participant 3 explains this through “I’m not interested in ads like of food or sports that I
don’t like. It doesn’t make sense to an advertisement for a beer because I don’t drink. I mostly see ads that I’m okay with.” Participant 1 supports this by stating “I know my data is being mined, which is unavoidable so I might as well get some recommendations and
advertisements that are good.”
The relevance of the content that consumers encounter plays a significant role, not only that it gives the consumers the ability to consume information that they might be interested in but also help them to absorb more information. This could be in terms of substitutes or
alternatives or a new product in general. Authors also dissect more on personalized advertisements in the following two main themes; change in requirement and shift in behavior.
4.1.2. Change in mode of advertising
When interviewees were asked about their preferred mode of advertising majority of the participants leaned toward the new AI-based advertisements rather than the old conventional way of advertising.
Participant 7 explains: ”I like a new mode of advertising; it certainly has a lot of potentials if
it’s used correctly.”
internet one way or the other and the old modes of availing information and entertainment are not as much used.
Participant 7 elaborate:
“We need to adapt to the new way of advertising because we don’t watch TV anymore, we’re different consumers now than we were 10 or 5 years ago.”
Participant 2 also explains their thoughts on the matter: “I certainly think that the new way of doing advertising has a lot of potentials, they make some of the greatest ads now.”
Not only that the participants favoured the new mode of advertising they also explain that the new mode of advertisement is more relevant but also less annoying and repetitive.
Participant 4 mentions: “I remember getting annoyed by most of the ads that would always be
the same and repeating over and over again, especially when I would be watching a game, but now I watch games online, I do get ads but they’re less repetitive and more about something that I would like.”
It can be seen through the responses of the participants that there was a much-needed gap in the advertisement industry that needed to be addressed. The transition of the advertising industry was much appreciated by the majority of the participants.
4.2. Changes in Requirements
The authors identify the second main theme to be changed in the requirement, based on the data collected. It became evident that most of the interviewees come across most of the advertisements through social media or any other online-based platforms. As the lifestyle of the respondents is continuously changing their requisites for consumption have also
transformed. These transformations can be categorized by personalized advertising, time efficiency, and a basic understanding of the technology (AI).
4.2.1. Personalized Advertisements
As mentioned above this can be seen as a reoccurring theme by the authors. However, here it is more of a practical implication. The majority of the participants require more tailored
advertisements or suggestions that interest them rather than something more focused on the general public. Although they still want some sort of transparency on how, where, and why their data is being mined and used.
Participant 9 states: “I find it a bit scary that you can sit and talk about something completely
random. And your phone is nearby, and you start receiving ads on it.”
Participants 5 and 7 also express their concern on the matter. They appreciate the accuracy of the content they receive but also feel skeptical about the idea of their phone listening to their conversations and receiving advertisements based on the topic of their conversation.
Participant 7 gives an example: “I was talking to a friend of mine about purchasing a new
computer chair, and I don’t play games, or I don’t look any of this stuff up, it was just
completely random. I didn’t make any searches, a few hours later while I was scrolling, I got ads for the gaming chair of this exact company we were talking about.”
Upon asking how it makes them feel as a consumer when their data is used for tailored advertisements and how it affects their relationship with the brand the answers were mostly positive. A lot of skepticism was raised by most respondents on how their data is used for them to receive these tailored advertisements. However, respondents also mention that its pros outweigh the cons.
Participant 2 express their sentiment as: “I know that a lot of people don’t like it, but I like it.
It would be weird to see ads for things that you are not interested in. So, I like it, sometimes you get new ideas, like things that you didn’t know existed. But I wouldn’t mind knowing how they know so much about me.”
This shows clearly that participants are concerned about their privacy however they do like the feature of receiving more personalized advertisements regardless of lack of privacy. More quotes related to personalized advertisements were mentioned above that clearly show the significance and importance. However, the authors felt that for the relevance of this paper needed to use these quotes where they feel the most relevant.
4.2.2. Time Efficiency
Time efficiency was an emerging theme by the participants. The participants showcased a specific interest in the duration of advertisements that they receive and the time they save through recommendation systems while they are browsing or shopping online. Participants mentioned that they save an ample amount of time especially while shopping online. Participant 6 mentions that: “I shop online for almost everything even groceries and food. I
feel shopping online saves me a lot of time. For example, if I’m buying a pair of pants often finding something to pair with it through the recommendation is a bit more convenient than searching it by myself.”
Participants also suggested the duration of the advertisements plays a crucial role to capture their interests. Shorter advertisements are able to deliver their message directly have a long-lasting effect on them and grab their attention better than longer advertisements. According to participants 3 and 10, the advertisements that they come across on social media or YouTube that have a runtime of few seconds are received as more interesting. Participant 3 quotes:
“Yesterday on Instagram I saw an advertisement about this cola called Cuba Cola, it was funny and very short. I think this is what I would categorize as a good advertisement.”
Similarly, the participant states: “In my opinion, an ad should be able to deliver its message
within the first 5 seconds, that’s when you grab the most attention, in my opinion”
4.2.3. Basic Understanding of AI
In the interview conducted by the authors, the realization was soon made that almost all of the participants were quite familiar with AI and its field of expertise. Most of the participants involved had an almost clear view of what AI is, what it has to offer, and what could be the potential of machine learning in the future.
The participants were asked a series of questions regarding the use of machine learning within the online marketing field and how it might affect their purchases and decision-making process.
Participant 1: this individual was very familiar with AI in general and had a general
mind is digital transformation. And that is because AI plays a major part within digital transformation and especially the digital transformation within different organizations.”
He later added that the use of machine learning and AI helps companies to reach a bigger target audience.
Participant 2: this interviewee had a more fiction-like understanding of AI even though he fairly understood the capabilities of machine learning.
He went on to state: “when it comes to AI, I think of like robots and machinery that are
aware of their surroundings”.
He went on to explain that he thinks that AI is at an ever-growing rate and it’s not just about robots and would eventually turn into something that human beings can heavily rely on to carry out day-to-day tasks.
Participant 3: another subject to this interview was a female studying in the field of ICT and had a more technical and deeper understanding in general about AI. She mentioned that: ”I
know that AI is being used in marketing today, especially with algorithms made by big companies such as Facebook”
4.3. Shift in Behaviour
After carefully evaluating the interviews, the authors quickly concluded that almost all of the participants had a clear understanding of AI and its uses within marketing and online
marketing.
4.3.1. Personalized Advertisement
On the other hand, most of the participants admitted that they find these ads very beneficial to them and it helps them get to what they need easier. One of the participants mentioned:
“sometimes I would look for a product but have no luck finding it and later on I would come across it on a social media ad and purchase it”.
4.3.2. Changes in Purchase Behaviour
Another question that came to mind, however, was whether the application of machine learning in online marketing affects the purchase behavior of consumers or not.
In this scenario, the interviewees had different answers and justifications at times. Some believed that it doesn’t have a direct effect on their shopping behavior. For example, if one of the participants said that they only shop if they need something and don’t participate in stress shop or seldom get tempted by ads they see online. They added on by saying that they would actively look for the goods they need and find these ads as sort of traps for companies to get consumers to buy their products.
4.3.3. Attitude Towards AI
Overall, most of the people who were interviewed by the authors had a positive attitude towards the use of AI within marketing and found it very useful. Even though there were some questions raised when it came to the safety and security aspects of it. The most serious of the questions was whether or not the data collected by these machines were in safe hands. They questioned the moral side of the story too. Some of the participants in the interview, came up with some ideas to try to help this gray idea of AI in marketing. One of the participants added
“it would be a lot better if it was made clear to users which data they were collecting from them and how were they collecting these data”
5. Analysis
This section analyses the findings that were obtained from the primary data. Where it aims to answer the research question of this thesis: How does AI in marketing and advertising
influence consumer's decision behavior? Primary data combined with secondary data from
academic articles and theories will be presented to provide more generalized findings. The analysis in this section with the help of the main themes and their underlying sub-themes has specifically influenced the consumption of the participants through AI-based advertisements in the context of Blackwell et al. (2006) CDP model. These main themes extracted give an advantage to the following paragraphs to delve into the influence of AI technology in advertisements on change in consumer behavior.
The change in need was the first main theme that was identified by the authors, as the majority of the participants mention that the older traditional mode of advertising was something they don’t find attractive and want more personalized advertisements that are targeted specifically towards them and seeing something that they might be interested in. This fulfills the first criteria of the CDP model by Blackwell et al. (2006). According to the first main element, the participants have identified a need. The need in this being the preference of more relevant advertisements towards the consumers. As mentioned above by Wu and Wen (2021), consumers appreciate AI-created advertisements, and consumer's perceived
objectivity is also influenced positively towards machine heuristics. In short, consumers believe that the advertisements that are generated through machine learning (AI technology) are more trustworthy.
Not only that the consumer appreciated the personalized advertisements they also mention that how necessary it was to change the mode of advertising. Most of the consumers stated that due to this they can see more quality content which they don’t see on the older
conventional platforms such as TV. This is because nowadays consumers spend more time on digital devices and based on what they have searched or looked up before they get content of their interest. As stated above AI technology evaluates the real-time scenarios by collecting data from physical or digital sources and provides personalized recommendations, solutions, and alternatives to consumers to solve their problems. Not only that AI has made it possible to engage more consumers through interactive advertising. This has empowered consumers to provide direct feedback, this enables marketers and organizations to further improve their
marketing and advertising efforts, such as market research and analysis, design, and
copyrights. Consumers welcomed the ability to be able to get their opinions heard as a much-needed change in the advertising industry.
The second main theme talks about how the requirements have changed for consumers. As the majority of the participants treasure tailored advertisements because now they don’t have to wait to know about a new product or service or actively search for it. The amount of time that they spent on searching is now significantly less. Meaning that now they can find more information in less time. This satisfies the second element of the CDP model; search for information. As mentioned above, Taulli (2020) states that programmatic advertising has fashioned an environment where it is easier to retain efficiency in message creation and delivery. Because of this conceiving and being exposed to personalized brand communication is reasonably higher. People think cognitively and usually chose options that are effortless to think and solve problems instead of effortful and sophisticated ones (Fiske and Taylor, 1991). This could be associated with the findings of Wu and Wen (2021), that people look for an easier way to take mental shortcuts to process information and avoid effortful ways of analyzing information.
Personal advertisements are categorized as one of the categories in this theme that plays a crucial role. Even though, the consumers appreciated the relevancy of the advertisements as per their taste they required more transparency on the matter of their data being collected. Also, the AI changes implemented by management are not always smooth and often or not create problems for the consumers to interact with the company (Ene, 2018). As previously mentioned by Hagendorff (2020) that there is a serious lack of infrastructure in ethics in AI implementation. These ethical weaknesses could be the issue of transparency as well. As previously stated, steps are being taken by organizations themselves to counteract these problems and become more and more ethical with the implementation of AI.
However, the participants also mentioned that despite there is a lack of transparency they like the idea of tailored advertisements as they get to see different alternatives, products, and services that they were previously not aware of. This fulfills the third step of the CDP model; pre-purchase evaluation. Consumers get to evaluate the best possible option that will fulfill their needs the most. Brands have started to adopt AI to promote their products and provide
language and image generation, they learn what consumers behavior and suggest products that are per their taste. This not only increases the effectiveness of the tailored advertising but also provides the consumer with a refined experience.
The subjects interviewed and analyzed throughout the research, mostly provided the same approach and answers towards how AI implemented in marketing affected their behavior as consumers. All the subjects showed a decent understanding of AI and machine learning in general. As mentioned above, almost all of the participants had a clear view and explained their basic understandings towards their understanding of AI.
Furthermore, sets of questions were asked by the authors regarding the effects on purchase behavior within the targeted market which consists of consumers. Most of the subjects to the interviews believed that their purchasing behavior as consumers was directly affected due to the existence of machine learning.
Most of the participants were amazed by how much data is being collected from them and how they randomly get tailored ads that show them exactly what they were looking for without them actively searching for those items.
The conclusion could almost be drawn from the evidence collected that AI and consumer behavior offers good insights when closely observed and implemented together. It is a rather complex task to gather evidence and analyze how consumers are shopping these days due to the availability of a large amount of data. This was agreed upon by the subjects to the interviews as mentioned above.
Another main topic discussed by the interviews was the change accrued in the main theme that was mentioned by the authors. Most of the interviewees found the old traditional advertising method out fashioned and boring since it couldn’t satisfy the needs of the consumers and the new generation of people in the new era would prefer something more personalized and tailored in a way that would satisfy their needs.
Furthermore, the consumers believed that AI has helped the advertising industry massively and has almost given it a new life altogether. Nowadays consumers have digital platforms in
much more content made accessible through social media and online platforms through AI than ever before and it has made the advertising industry almost limitless.
6. Conclusion
The empirical findings and the analysis suggest the aim to answer the research question of this thesis: How does AI in marketing and advertising influence consumer's decision behavior? After conducting in-depth research through reliable sources and the primary data collected through interviews combined with secondary data from articles, it has become evident how the existence of AI in marketing influences consumer behavior. Through the scope of Blackwell et al. (2006) CDP model, the collected data has been analyzed and the results give an advantage regarding the influence of AI technology in advertisement on change in consumer behavior.
One of the main themes identified, was the change in need. After the evaluation of the interviews, the authors quickly identified that the bulk of the participants mentioned that the traditional way of advertising has outgrown them and has become boring, and they prefer advertisements that are tailored in a way that they are personalized with their interests and likings. This aligns with the first element in which the participants in the interview have recognized a need. The need in this has been identified as wanting more directly targeted advertisements.
As mentioned above in the analysis, consumers went on to mention how important the
change in settings to the advertisements has been. This has caused more consumers to be able to have access to the ads at almost any given time in any location. That is caused by the fact that technological advancements have allowed us to have access to the database on a variety of digital devices hence smoothening the path for online advertisements. As mentioned previously it cuts the duration of information being searched by the consumer in addition to that making the process of evaluating different products and services prior to purchase easier.
Artificial intelligence helps marketers and companies collect data from consumers through the things they look up and search online. This opens a path towards trust issues between consumers and machine learning. As mentioned by Hagendorff (2020) there is a massive lack of ethical values within the deployment of AI and the use it. As mentioned by the subjects to the interview, they appreciate AI and the way it has changed the perception of how
