No products in the cart.
6.2. Landmark Publications
Next, we identify the landmark publications that have contributed to form and
grow research on AI in marketing. For this analysis, we consider betweenness
centrality. It is a measure of centrality in a network based on the shortest paths, such
that any node with higher betweenness centrality exerts more influence over the
network because more information passes through it, hence it exerts a high influence.
All the nodes in a network can be assigned relative scores, with the predictions that
connections to other, high scoring nodes contribute more to the focal node’s score than
do equal connections to low scoring nodes. Thus, a higher score means that a node is
connected to multiple other nodes that have high scores themselves.
For our study, rather than simply counting the number of citations of the articles
in the reviewed literature, we seek to identify the centrally located documents on the
basis of co-citation patterns, because they are the articles that bind the research field.
For example, Barney’s (1991) article—Special theory forum the resource-based model
of the firm: origins, implications, and prospects—is not part of our literature pool, but
exerts high influence on in the extant studies on AI in marketing. For this analysis, we
apply algorithms that measure how often two articles are cited together by a third, as
well as how often a pair of articles cite the same third article. In both cases, the premise
is that paired articles should share knowledge commonalities. In Table 5, we list the
top 10 articles in terms of their centrality scores.
Table 5: Significant publications on AI in marketing in terms of centrality
6.3. Density Visualization
For this scientometric analysis, we conduct a keyword density analysis, which
identifies words and concepts that are most prominent in the specific field. In Figure 8,
larger bubbles imply the more frequently used keywords and associated concepts; for
this density visualization, we set a minimum threshold of at least five occurrences in the
literature pool. It also reveals the location trends of dense point data, with color
gradients, such that we can identify where groups of keywords and their associated
concepts correlate
Starting from the top left corner, we find that studies with an organizational
perspective represent a dense area. They examine big data and their impacts on
organizational practices, performance, competitive advantage, and so on. Moving
clockwise, we find a density cluster pertaining to superior value creation for
consumers through AI, involving various forms of information, risks, and innovation.
Then we note multiple dense keyword clusters formed by studies that focus on, for
example, IoT, web, and data mining. At the bottom, analytical studies address multiple
variables, along with Bayesian networks, loyalty, nature, and RFM (Recency,
Frequency, Monetary value). Finally, a cluster includes studies of computing systems
that resemble biological neural networks and how AI can be used to enhance the
abilities of marketers.
6.4. Keyword Co-Occurrence Analysis
Keyword co-occurrence networks are the mutual interconnectedness of terms
according to their paired presence within the literature base. In knowledge mapping,
each node represents a keyword, and each link represents the co-occurrence of a pair
of words. The weight of a link connecting each pair represents the number of times
these words co-occur in multiple articles. Therefore, the co-occurrence network
effectively represents the cumulative knowledge of a domain, in terms of its crucial
knowledge components and insights, as established by patterns and strength of links
between keywords that appear in the literature (Radhakrishnan et al., 2017; Wu et al.,
2019).
In Figure 9, depicting the keyword co-occurrence network for AI in marketing,
the lower-left corner pertains to studies on neural networks and their tight connections
with studies of market segmentation, model development, and behavioral research.
Similarly, we find strong co-occurrences across studies that focus on sales, WOM, user
generated content, and quality. In the center, the highest co-occurrence values link
“big data” with “model,” which in turn are connected with a multiplicity of other
nodes. Performance and impact considerations are central too, reflecting the outcome
oriented nature of these studies. On the right side, keyword co-occurrences indicate a
strong presence of AI studies linking big data analytics with capabilities, competitive
advantage, or firm performance. Then at the bottom, we find a strong presence of
studies that connect firm management and strategy with data mining, analytics, and
innovation through AI.
Figure 9: Keyword co-occurrences
6.5. Mapping Central Keywords over Time
With a further examination of the keywords, we investigate the progression of
central focus areas over time. These analyses and subsequent plotting of keywords
were performed by algorithms in Citespace. They reveal that the research field has
become more diversified in recent years, so we concentrate on multiple keywords in
the similar time periods, which become apparent through their overlaps, as we depict
in Figure 10. In particular, the dynamic evolution of keywords indicates a shift in focus,
from information and methods (text mining) to more defined applications of AI
techniques to marketing problems. The keywords appearing after around 2015
highlight various aspects of AI in marketing, such as its influence on customer
satisfaction, WOM, or firm performance. They also signal pertinent challenges, such as
the use of big data and analytics for gaining competitive advantage or the importance
of building AI capabilities within the firm.
Figure 11 depicts scientific collaborations among the leading researchers in the
field, identifying three prominent researchers in the most central positions: Hein Ruys,
Jaesoo Kim, and Sherrie Wei. Their work connects to other major researchers too,
including (on the left) Roland Schegg, Paul Phillip, and Maria Manuela Santos Silva and
(on the right) Alfonso Palmer, Juan José Montaño, and Albert Sesé. Many of these
researchers study uses of AI in tourism marketing.
In addition to researchers, the source journals symbolize the domain. Exploring
and reviewing the underlined publication trends among leading journals is likely to
significantly help any research institution to identify the core strengths of the
intellectual domain and reinforce scientific cooperations. Moreover, the analysis
reveals the formation and development of intellectual networks within the knowledge
base and provides valuable insights into the journals that have collectively shaped the
field. In Figure 12, the co-citation network among leading journals that published
articles on AI in marketing shows the prominence of Journal of Marketing and Journal of
Marketing Research, which cite each other extensively. The Journal of the Academy of
Marketing Science, MIS Quarterly, Harvard Business Review, Journal of Business Research,
and Strategic Management Journal complement the core of this research domain
6.2. Landmark Publications
Next, we identify the landmark publications that have contributed to form and
grow research on AI in marketing. For this analysis, we consider betweenness
centrality. It is a measure of centrality in a network based on the shortest paths, such
that any node with higher betweenness centrality exerts more influence over the
network because more information passes through it, hence it exerts a high influence.
All the nodes in a network can be assigned relative scores, with the predictions that
connections to other, high scoring nodes contribute more to the focal node’s score than
do equal connections to low scoring nodes. Thus, a higher score means that a node is
connected to multiple other nodes that have high scores themselves.
For our study, rather than simply counting the number of citations of the articles
in the reviewed literature, we seek to identify the centrally located documents on the
basis of co-citation patterns, because they are the articles that bind the research field.
For example, Barney’s (1991) article—Special theory forum the resource-based model
of the firm: origins, implications, and prospects—is not part of our literature pool, but
exerts high influence on in the extant studies on AI in marketing. For this analysis, we
apply algorithms that measure how often two articles are cited together by a third, as
well as how often a pair of articles cite the same third article. In both cases, the premise
is that paired articles should share knowledge commonalities. In Table 5, we list the
top 10 articles in terms of their centrality scores.
Table 5: Significant publications on AI in marketing in terms of centrality
6.3. Density Visualization
For this scientometric analysis, we conduct a keyword density analysis, which
identifies words and concepts that are most prominent in the specific field. In Figure 8,
larger bubbles imply the more frequently used keywords and associated concepts; for
this density visualization, we set a minimum threshold of at least five occurrences in the
literature pool. It also reveals the location trends of dense point data, with color
gradients, such that we can identify where groups of keywords and their associated
concepts correlate
Starting from the top left corner, we find that studies with an organizational
perspective represent a dense area. They examine big data and their impacts on
organizational practices, performance, competitive advantage, and so on. Moving
clockwise, we find a density cluster pertaining to superior value creation for
consumers through AI, involving various forms of information, risks, and innovation.
Then we note multiple dense keyword clusters formed by studies that focus on, for
example, IoT, web, and data mining. At the bottom, analytical studies address multiple
variables, along with Bayesian networks, loyalty, nature, and RFM (Recency,
Frequency, Monetary value). Finally, a cluster includes studies of computing systems
that resemble biological neural networks and how AI can be used to enhance the
abilities of marketers.
6.4. Keyword Co-Occurrence Analysis
Keyword co-occurrence networks are the mutual interconnectedness of terms
according to their paired presence within the literature base. In knowledge mapping,
each node represents a keyword, and each link represents the co-occurrence of a pair
of words. The weight of a link connecting each pair represents the number of times
these words co-occur in multiple articles. Therefore, the co-occurrence network
effectively represents the cumulative knowledge of a domain, in terms of its crucial
knowledge components and insights, as established by patterns and strength of links
between keywords that appear in the literature (Radhakrishnan et al., 2017; Wu et al.,
2019).
In Figure 9, depicting the keyword co-occurrence network for AI in marketing,
the lower-left corner pertains to studies on neural networks and their tight connections
with studies of market segmentation, model development, and behavioral research.
Similarly, we find strong co-occurrences across studies that focus on sales, WOM, user
generated content, and quality. In the center, the highest co-occurrence values link
“big data” with “model,” which in turn are connected with a multiplicity of other
nodes. Performance and impact considerations are central too, reflecting the outcome
oriented nature of these studies. On the right side, keyword co-occurrences indicate a
strong presence of AI studies linking big data analytics with capabilities, competitive
advantage, or firm performance. Then at the bottom, we find a strong presence of
studies that connect firm management and strategy with data mining, analytics, and
innovation through AI.
Figure 9: Keyword co-occurrences
6.5. Mapping Central Keywords over Time
With a further examination of the keywords, we investigate the progression of
central focus areas over time. These analyses and subsequent plotting of keywords
were performed by algorithms in Citespace. They reveal that the research field has
become more diversified in recent years, so we concentrate on multiple keywords in
the similar time periods, which become apparent through their overlaps, as we depict
in Figure 10. In particular, the dynamic evolution of keywords indicates a shift in focus,
from information and methods (text mining) to more defined applications of AI
techniques to marketing problems. The keywords appearing after around 2015
highlight various aspects of AI in marketing, such as its influence on customer
satisfaction, WOM, or firm performance. They also signal pertinent challenges, such as
the use of big data and analytics for gaining competitive advantage or the importance
of building AI capabilities within the firm.
Figure 11 depicts scientific collaborations among the leading researchers in the
field, identifying three prominent researchers in the most central positions: Hein Ruys,
Jaesoo Kim, and Sherrie Wei. Their work connects to other major researchers too,
including (on the left) Roland Schegg, Paul Phillip, and Maria Manuela Santos Silva and
(on the right) Alfonso Palmer, Juan José Montaño, and Albert Sesé. Many of these
researchers study uses of AI in tourism marketing.
In addition to researchers, the source journals symbolize the domain. Exploring
and reviewing the underlined publication trends among leading journals is likely to
significantly help any research institution to identify the core strengths of the
intellectual domain and reinforce scientific cooperations. Moreover, the analysis
reveals the formation and development of intellectual networks within the knowledge
base and provides valuable insights into the journals that have collectively shaped the
field. In Figure 12, the co-citation network among leading journals that published
articles on AI in marketing shows the prominence of Journal of Marketing and Journal of
Marketing Research, which cite each other extensively. The Journal of the Academy of
Marketing Science, MIS Quarterly, Harvard Business Review, Journal of Business Research,
and Strategic Management Journal complement the core of this research domain