Artificial Intelligence Applications on the Influencing Factors of
Short Video Marketing on Consumers' Buying Behavior
Yixuan Wu
University of Southern California, U.S.A.
Keywords: Consumer Psychology, Purchase Decision, Artificial Intelligence, Natural Language Processing,
Reinforcement Learning.
Abstract: Live streaming has emerged as an emerging marketing strategy that has a significant impact on consumers'
purchasing psychology and behaviour as human society transitions to the mobile Internet era. This is espe-
cially true in the wake of the recent epidemic and the domestic consumption stimulation cycle, where live
streaming has become a new consumption trend. In this study, a model of the impact of shared recommenda-
tions on customers' purchasing decisions is constructed based on two factors: product display vividness and
user engagement, which is based on the idea of computer-mediated communication. The findings of the ex-
periment demonstrate that consumer attention is directly influenced by product vividness and user interaction,
both of which have a significant impact on consumers' psychosocial distance and emotional intensity, and
consequently have a significant influence on consumers' purchasing decisions.
1 INTRODUCTION
Since the dawn of the twenty-first century, human-
kind has greatly improved material production, infor-
mation and communication technology (ICT) and
mobile Internet technology have advanced quickly,
and the role of emerging media in the mobile Internet
era has grown in importance in people's daily lives.
Statistics show that China has the most Internet users
in the world, with close to 1 billion users as of March
2021. Of these, more than half are using mobile de-
vices, and more than 700 million people utilize online
shopping sites (Bygstad, 2016).
Consumers are growing weary of the novelty of
the e-commerce model and are dissatisfied with the
lack of new developments in e-commerce shopping.
In order to address the varied wants of consumers,
new marketing strategies are required. As a result,
creative e-commerce marketing models have
emerged as the key to overcoming the industry's de-
velopment bottleneck.
The premise and model of this study are based on
the current theoretical research on live-streaming and
in-depth theoretical analysis. This study is an in-depth
investigation of the marketing phenomena of live-
streaming in the new media in the context of the mo-
bile Internet. Through qualitative interviews, ques-
tionnaire surveys, structural equation modelling, and
simulation tests, this study explores how the live-
streaming marketing technique influences customers'
psychology in the new media environment and how
this influence ultimately affects consumer behavior.
2 LITERATURE REVIEW
On mobile social commerce platforms like Mush-
room Street and Xiaohongshu, there are currently a
huge amount of sharing recommendations made by
regular users that successfully pique people's interest
in purchasing things as well as promoting and selling
them. In contrast to conventional e-commerce plat-
forms, it primarily uses videos to present products and
provide consumers with pertinent product infor-
mation as opposed to text descriptions and pictures.
Short videos are also used to convey the effects and
feelings of using products while fostering a certain
level of interactive experience. China's social e-com-
merce market reached 2 trillion in value in 2019, a
notable rise of 71.71% from the previous year; social
e-commerce users increased by 17.26% to 713 mil-
lion, representing a sizable business opportunity
(China Internet Association, 2019). The platform's
458
Wu, Y.
Artificial Intelligence Applications on the Influencing Factors of Short Video Marketing on Consumersâ
˘
A
´
Z Buying Behavior.
DOI: 10.5220/0011740000003607
In Proceedings of the 1st International Conference on Public Management, Digital Economy and Internet Technology (ICPDI 2022), pages 458-462
ISBN: 978-989-758-620-0
Copyright
c
2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)
activity and product relevance have grown dramati-
cally as a result of the tight user contact, and it is now
better able to satisfy the customers' increasingly cus-
tomized and varied consuming needs (China Social
E-Commerce Market Data Report 2019).
2.1 Reinforcement Learning
The boundaries of time and geography have been fur-
ther dismantled by information interchange and social
contact, and people's acceptance, recognition, and en-
gagement in online product recommendations
(OPRs), particularly those provided by regular con-
sumers, is growing. In conclusion, social commerce
is quickly developing into a scalable and high-growth
sector of the m-commerce business, representing an
important trend in the future development of e-com-
merce, thanks to the ongoing optimization and inven-
tion of online product suggestions. It is crucial to ob-
tain insight into the psychological conditions and be-
havioral goals of consumers' online consumption in
order to comprehend the efficiency of shared recom-
mendations.
2.2 Natural Language Processing
According to Hajli and Sims (2015), social commerce
may also refer to a social platform that came into ex-
istence with the rise of Web 2.0 and allows users to
create original content and share their experiences.
their encounters (Dartmouth College. 2017). This
study summarizes and investigates the impact of
shared recommendations on consumers' purchase de-
cisions at the psychological level based on the afore-
mentioned theoretical underpinnings and the analysis
of the current research situation, then proposes the re-
search model, as shown in the Figure 1. Natural lan-
guage processing, which is the study of theories and
techniques for successful human-computer interac-
tion using natural language, is one of the most chal-
lenging problems in the field of artificial intelligence.
2.3 Eye-Movement Experiment
Our focus has always been finite, valuable, and hard
to come by. But what makes the present different is
that technology advancements have made a dizzying
amount of information available that is purposefully
meant to grab our attention. It has never been simpler
for the general population to attract such high
amounts of personal attention via platforms like so-
cial media.
Eye tracking methods fall under the category of
Neuro-Physiological (NP) approaches, which use oc-
ulomotor devices to record people's eye movements
in a specific environment. The data from this combi-
nation is then used to analyze and discuss the inherent
relationship between eye movements and cognitive
activity. Around 80% to 90% of external information
is acquired through the eyes, and the human brain sig-
nificantly relies on vision for information processing
(Hong, 2020). The direct hypothesis and the eye-brain
hypothesis of American psychologists Just and Car-
penter both indicate to a direct connection between
eye movements and cognition, and cognitive psychol-
ogy has thus discovered many reliable connections
between eye movements and cognition (Ian, 2001).
2.4 Eye Movement Indicator Meaning
and Interpretation
In order to understand the peculiarities of participants'
information search and browsing activities, gaze - an
important aspect of eye movements - is analyzed.
Information processing and gaze are strongly con-
nected. For a brief moment as a person gazes, their
eyes are relatively still as their brain extracts infor-
mation. To analyze reading behavior patterns, page
elements, page design recognizability, etc., eye
movement metrics, such as gaze duration, number of
gaze points, average pupil size, and other common fo-
cus metrics, can be employed.
Figure 1: Research Model (Drafted by author).
Artificial Intelligence Applications on the Influencing Factors of Short Video Marketing on Consumersâ
˘
A
´
Z Buying Behavior
459
Figure 2: Eye-movement experiment diagram.
Figure 3: Flow chart of eye tracking scenario experiment (Drafted by author).
2.5 Computer Technology and
Consumer’s Behavior
With the usage of streaming video and audio, online
advertising is starting to resemble television advertis-
ing, the most widely used form of media. Conse-
quently, this kind of media may be due to its enter-
tainment value, be more appealing to Generation Y
consumers. The article mentions that MSN, Yahoo!,
and AOL websites are the most widely used platforms
for internet advertising.
Internet users conduct research and add to the data
they obtain through conventional advertising. As a re-
sult, advertisements that direct viewers to a useful in-
ternet resource are more likely to succeed than those
that don't. According to the survey, television, radio,
magazine advertising, direct mail, and billboard cam-
paigns are those that have been most positively im-
pacted by the Internet.
3 METHODOLOGY
This study is grounded in the marketing theory
AIDMA model (attention - interest - desire - memory
- action). This study explores how live streaming af-
fects consumer psychology and behaviour in the con-
text of new media, as well as how it works. In partic-
ular, the interactivity, entertainment, content, anchor
appeal, trust, and fit of live streaming in the new me-
dia ecosystem are considered independent variables
in this study. Consumer perceptions of a brand's value
and buy intention are considered intermediate varia-
bles, and actual consuming behavior is considered the
dependent variable (Luo, 2013). The eye-tracking ex-
periment's particular data and analyses are included
in this thesis along with the hypothesis testing and
testing process.
3.1 Methodology Design
Qualitative interviews, questionnaires, and simulated
experiments served as the foundation for this study.
For analysis and statistical testing to guarantee the
study's conclusions are factual and scientific, qualita-
tive interviews, questionnaires, and simulated exper-
iments were employed.
3.2 Eye Movement Experiment Data
Collection
The EyeLink 1000 plus creates a separate EDF file
with each subject's complete experimental data and
saves it on their computer at the conclusion of the eye
tracking trial.
The DataViewer software is used to view the eye-
movement data files, group the data into time periods
for all valid subjects, slice the data into those time pe-
riods, create gaze hotspot maps, and export the nec-
essary eye-movement indicators, such as total gaze
duration, total number of gaze points, first gaze time,
first gaze duration, mean pupil size, and so forth.
ICPDI 2022 - International Conference on Public Management, Digital Economy and Internet Technology
460
Figure 4: Survey question extraction and sources (Drafted by author).
Figure 5: Results of regression analysis of consumers' psychosocial distance and their emotional intensity (Drafted by author).
3.3 Hypothesis Testing and Analysis of
the Relationship between
Psychosocial Distance and
Emotional Intensity
The findings that consumer psychosocial distance has
a substantial positive effect on affective intensity (1=-
0.647, p1=0.0000.001; 2=-0.663, p2=0.0000.001)
support the hypothesis H2a. According to the experi-
mental findings, customers' emotional intensity in-
creases with increasing psychosocial distance from
them. First, the significance level was met at the
=0.001 level and the F-statistic of psychosocial dis-
tance in the two experimental contexts with different
levels of vividness of product display was
p=0.0000.001, indicating that psychosocial distance
is sufficient to explain the change in emotional inten-
sity. Third, as a general guideline for covariance as-
sessment, the variance inflation factor (VIF) value
created for each predictor variable must be less than
5 [186], and the VIF values were around 2 for both of
the various experimental stimulus settings, showing
no autocorrelation. The absolute value of the stand-
ardised regression coefficient, which has beta values
of -0.647 and -0.663 in the two contexts of product
display vividness, respectively, represents the degree
of effect of the independent variable on the dependent
variable. The larger the absolute value of the stand-
ardised coefficient, the greater the degree of influ-
ence. The Beta values are -0.647 and -0.663 for the
two product display vividness scenarios, indicating
that psychosocial distance negatively affects affective
intensity, and consumers perceive that the effect of
psychosocial distance on their affective intensity in-
creases when the product display is more vivid.
4 DISCUSSION
The acceptance of short video marketing, a recently
developed marketing strategy, varies depending on
the region, the gender, the age, and the income level
of the community. It is important for businesses to use
unique marketing strategies for various customer seg-
ments to increase marketing's focus, which is better
Artificial Intelligence Applications on the Influencing Factors of Short Video Marketing on Consumersâ
˘
A
´
Z Buying Behavior
461
for the growth of marketing initiatives and marketing
goals. Therefore, in future research, these elements
can be examined as variables (Trope, 2010).
It is worthwhile to investigate which products are
more appropriate for short video marketing and which
products are more effective through short video mar-
keting because there are many products that can be
sold through short video marketing and different
products have different characteristics, use cycles,
and people (Turvey, 2000). The issue merits investi-
gation. In order to determine the most successful mar-
keting strategies for linked businesses, future study
can be divided into product kinds and explore the var-
ious effects of various items in short video marketing.
The variable of psychological distance has been
studied extensively and there are various views on the
dimensional classification of psychological distance.
At present, the more accepted dimensional classifica-
tion in academic circles is divided into three dimen-
sions: temporal, spatial and social.
Prior to summarizing and improving the availabil-
ity of short-form video marketing, the availability re-
search in the areas of social commerce and online
shopping is analyzed. In short-form video marketing
scenarios, consumers' purchasing decisions can be in-
fluenced by both feature and platform accessibility.
According to the prior empirical investigation, plat-
form availability has a greater influence overall in
terms of strength. The impact of feature availability,
however, is a little weaker. The purpose of computer
vision technology is to extract information from ob-
jective images, interpret them, and comprehend them
by simulating human visual functions on computers.
At this time, content monitoring, intelligent logistics,
and product picture search are the key contributions
of computer vision technology to e-commerce.
Customers are more able to promote their pur-
chase behaviour in the new consumer environment of
short-form video shopping due to their increased in-
terest in new forms of technology, communication,
and interaction, especially the algorithmic recom-
mendation mechanism of short-form video, which has
a higher accuracy of distribution, a benefit that tradi-
tional online shopping does not have.
REFERENCES
Bygstad B., Munkvold B. E., Volkoff O. Identifying Gen-
erative Mechanisms Through Affordances: A Frame-
work for Critical Realist Data Analysis[J]. Journal of
Information Technology, 2016, 31(1).
Chi H. Interactive digital advertising vs. virtual brand com-
munity: Exploratory study of user motivation and social
media marketing responses in Taiwan [J]. Journal of in-
teractive advertising, 2011,12(1):44-61.
China Internet Association. 2019 China Social E-Com-
merce Industry Development Report - From Innovation
in Diversion Mode to Systematic Operation.
China Social E-Commerce Market Data Report 2019 [R].
Netscape E-commerce Research Center, 2020.
Dartmouth College. 2017. “Extraocular Eye Movements.”
http://www.dartmouth.edu/~dons/figures/chapt_4/Fig
_4-1.htm.
Horng S, Wu C. How behaviors on social network sites and
online social capital influence social commerce inten-
tions [J]. Information & Management, 2020, 57(2):
103176.
Ian H. Technologies, Texts and Affordances [J]. Sociology,
2001, 35(2).
Luo X, Zhang J, Duan W. Social Media and Firm Equity
Value [J]. Information Systems Research, 2013, 24(1):
146-163.
Trope Y, Liberman N. Construal-Level Theory of Psycho-
logical Distance[J]. Psychological Review, 2010,
117(3).
Turvey M T, Shockley K, Carello C. Affordance, Proper
Function, and the Physical Basis of Perceived Heavi-
ness[J]. Cognition, 2000, 73(2).
Wang Y, Yu C. Social interaction-based consumer deci-
sion-making model in social commerce: The role of
word of mouth and observational learning[J]. Interna-
tional Journal of Information Management, 2017,
37(3):179-189.
ICPDI 2022 - International Conference on Public Management, Digital Economy and Internet Technology
462