INTRODUCTION
The increasing
growth of technology that is in line with the development of the internet in
Indonesia is very helpful for the ease of socializing and interacting between
social beings (Winarso, 2020). Likewise in
Indonesia, where almost 75% of the total population uses the internet. As many
as 277.7 million of the total population in Indonesia, half of them are
internet users with a total of 204.7 million. This can be seen in Figure 1.
Likewise, with increasingly rapid technological advances, smartphone users in Indonesia
are also increasing, meaning that more and more cellular phones are connected
to the internet.
Figure 1 Indonesian Internet
User Data 2022
Source:
We Are Social Hootsuite (2022)
With the advancement in the
use of technology and the internet, which is increasingly high, many social
beings spend most of their time on social media. The We Are Social Hootsuite
(2022) data in Figure 2 shows that the time spent is eight hours and thirty-six
minutes for each day. With the use of social media that dominates compared to other internet uses.
Figure
2 Time of Media Use in Indonesia in 2022
Source: We Are Social Hootsuite (2022)
Explanations from Figures 1 and 2 explain that with the
advancement of internet use in Indonesia, the use of social media is also
increasing. So that more and more companies are using social media as a medium
to market their products, which is done on social media because it does not
recognize space and time, meaning that it can be done anytime and anywhere with
a wider market and consumer.
Instagram is the second highest widely used digital social
media platform, after whatsapp. Almost the entire
population in Indonesia is Instagram users, according to data reported by We
Are Social Hootsuite, 84.8% of the population in Indonesia has downloaded the
Instagram application, and the rest use Facebook, Tiktok,
Telegram, and many other social media. used by people in Indonesia. This is
because of the many interesting features provided by Instagram for its users
and has excellent image and video quality which can provide more value compared
to its competitors, which can only display one photo or video.
Figure
3. Platforms for Using Social Media
Source: We Are Social Hootsuite (2022)
The use of social media Instagram as one of the most used
social media platforms due to the uniqueness of Instagram in the display of
photos and videos is able to help marketers or a company in marketing their
products to the wider public without space and time, and can be accessed by
anyone.
Entering the middle of the year in August 2020, Instagram
again released its new feature called Reels, this feature had previously been
used in Brazil. The definition of reels itself is a short video that allows its
users to capture the moment within 15 seconds (www. Marketing.co.id). However,
the features presented on Reels have a duration of 1 minute, on Instagram Reels
the videos presented can be accelerated, slowed down, given sound effects and
music that can be adapted to the needs of the user, and have harmony in editing
or combine various clips for one video with very clear and good picture quality
(Maksimova & Savolainen, 2021).
Indonesia has become one of the countries with the most
users of the Instagram Reels feature since its launch, this was conveyed by
Country Director for Meta Indonesia, Pieter Lydian who stated that Indonesia is
the country with the most active Reels content creators. In addition, the use
of Reels in Indonesia (Zuharah & Tunggali, 2022). According to the
report quoted (www.cnnindonesia) within 5 months Reels has contributed for a
short time, with the state of Indonesia as the most users of Reels features.
The use process, and the way it works, which has the same
algorithm as Tiktok, has the advantage of being able
to edit content directly, making Reels one of the choices that marketers can
use to market their products on their Instagram social media accounts. The use
of product marketing carried out through Reels is in line with the statement Liu et al. (2019)
which states that the use of Reels can be done to save time, be faster in the
process of spreading to social media, and using Reels can also provide space
and time to potential consumers who They don't have much time to look for
information, and life necessities products, so they can still be up to date in
the search process on social media.
Since involvement of celebrity can affect consumer brand
attitude, meaning that if in the application of using Reels on Instagram some
variables that previously did not go through Consumer Brand Attitude as a
mediator variable (Wuisan et al., 2020). This study is
expected to see how the influence of which is positive and significant to the
consumer purchase intention variable, which will certainly help product
marketing activities on Instagram social media on the Reels feature. So that
from the activities carried out in this research, it is expected to be able to make a contribution that facilitates marketing activities
carried out by many companies and marketers, as well as making it easier for
prospective consumers to make purchases and review a product that is presented
through the Instagram platform.
Adam Mosseri
as Executive Director of Instagam stated that the use
of the Reels feature on Instagram will be more focused in 2022 due to changes
in very large technological developments, which also need to be improved on
supporting features on Instagram as well, one of which is Reels (Sharp & Gerrard, 2022). The feature
is focused as a feature on Instagram to monetize so that it can help all levels
of society, especially marketers to market their products, which is also
assisted by the involvement of several celebrities to facilitate the marketing
of these products.
METHOD
This study used a
quantitative research method. According to Sugiyono (2018)
quantitative research is a research method based on the philosophy of positivism,
used to examine certain populations or samples, data collection using research
instruments, quantitative or statistical data analysis, with the aim of testing
predetermined hypotheses. The population used in this study is the population
of internet users in Indonesia of 204.7 million with a percentage of Instagram
users of 84.8%, or almost all of internet users in Indonesia use Instagram.
Due
to the large number of populations, in this study the researchers used a
sampling of 100 respondents because according to Kriyantono (2020)
a good sample for research on two or more variables using path analysis or
SEM-PLS data analysis techniques is to use 100 samples. , so that in this study
100 respondents were used as research samples. The sampling technique used is
purposive sampling, meaning that the technique in selecting samples with
certain considerations, the use of purposive sampling is used because not all
samples can be used as appropriate criteria for research, so only respondents
use the Instagram application, and of course use the Reels feature that can be
sampled. Collecting data in this study using a questionnaire distributed via
google form by giving written statements to respondents to be answered. The
limit of the research carried out in this study starts from 6 July 2022 � 2
August 2022. With data processing techniques carried out using SEM with SMART
PLS tools.
Operational
Variables
Operational variables need to be done to
facilitate researchers in conducting research. According to Sugiyono (2018) the definition of variable operationalization
is an attribute of a person or object, or activity that has a certain variation
set by the researcher to be studied and then drawn conclusions. The above
definition can be locked that the operationalization of the variables needed to
determine the type, indicators, and scale of the variables involved in the
research, so that the test can be carried out correctly and in accordance with
the title of the study.
Table 1
Operational Variables
|
Variable
|
Indicator
|
Indicator
|
Variable Code
|
|
Interesting
Content (X1)
(Liu et al., 2019)
|
I'm always intrigued when I see short Reels videos
because they're so much fun
|
Questionnaire
|
IC1
|
|
I often watch Reels short videos because
they are so much fun
|
Questionnaire
|
IC2
|
|
I'll quickly understand something when I
see a short Reels video because it's so much fun
|
Questionnaire
|
IC3
|
|
Scenario-based
experience (X2)
(Liu et al., 2019)
|
Reels short video with lots of scenarios
in it makes it feel very real and appropriate
|
Questionnaire
|
SBE1
|
|
Most of the images in the short Reels
Video make me feel like I'm in the picture
|
Questionnaire
|
SBE2
|
|
Multiple consumption scenarios in short
videos enhance the expected consumption experience
|
Questionnaire
|
SBE3
|
|
The consumption of the scenario in the
short Reels video made me have a series of associations with expectations
|
Questionnaire
|
SBE4
|
|
User
Participation Interaction (X3)
(Liu et al., 2019)
|
I always give likes, comments, and share
content on Reels short videos
|
Questionnaire
|
UPI1
|
|
In interacting with outside audiences, I
feel more real and easy without space and time
|
Questionnaire
|
UPI2
|
|
I like to find information and improve
my cognitive ability by interacting with outside audience
|
Questionnaire
|
UPI3
|
|
When communicating with other people, I
tend to believe in the opinion of the majority
|
Questionnaire
|
UPI4
|
|
Perceived
Usefulness (X4)
(Davis,
1989)
|
With short videos Reels makes it easier
for consumers to find product searches
|
Questionnaire
|
PU1
|
|
Reels short videos support user
productivity
|
Questionnaire
|
PU2
|
|
When you want to find something
the search is done with Reels short video content is easy
|
Questionnaire
|
PU3
|
|
Can shorten the time to search for a particular
brand
|
Questionnaire
|
PU4
|
|
The features on Reels short videos are
very useful
|
Questionnaire
|
PU5
|
|
Using Reels video content provides a
lifestyle change for its users
|
Questionnaire
|
PU6
|
|
The features in the short Reels video
make it easy for me
|
Questionnaire
|
PU7
|
|
The use of short video Reels gives me
benefits and avoids something bad
|
Questionnaire
|
PU8
|
|
Perceived Enjoyment (X5)
(Baskara
& Sukaatmadja, 2016)
|
I feel comfortable during the whole
shopping process by just viewing a short video of Reels
|
Questionnaire
|
PE1
|
|
Watching various content in Reels short videos is
always very interesting
|
Questionnaire
|
PE2
|
|
By viewing the content on Reels short videos it
makes shopping easier for me
|
Questionnaire
|
PE3
|
|
I am very happy when I find a brand that
matches my personality through Reels' short videos
|
Questionnaire
|
PE4
|
|
Involvement
of Celebrity (X6)
(Ha
& Lam, 2017)
|
Following the daily activities of
celebrities is one of the most fun things for me
|
Questionnaire
|
IOC1
|
|
I really enjoy the activities that
celebrities do.
|
Questionnaire
|
IOC2
|
|
I like discussing celebrity activities with my
friends
|
Questionnaire
|
IOC3
|
|
When I participate in celebrity
activities I feel like I am myself
|
Questionnaire
|
IOC4
|
|
Educational activities carried out by
celebrities can make them my role models
|
Questionnaire
|
IOC5
|
|
Consumer
Brand Attitude (Z)
(Liu et al., 2019)
|
By watching a short Reels video, I will
remember a brand
|
Questionnaire
|
CBA1
|
|
By watching Reels short videos, I have a
new understanding of a brand
|
Questionnaire
|
CBA2
|
|
By watching Reels short videos, I have
positive feelings for the brand
|
Questionnaire
|
CBA3
|
|
Watching a short Reels video will
provoke my urge to buy a brand
|
Questionnaire
|
CBA4
|
|
I often introduce brands I know through
short Reels videos to my family and
�my friend
|
Questionnaire
|
CBA5
|
|
Consumer Purchase
Intention (Y)
(Ha
& Lam, 2017)
|
I intend to make a purchase on each
brand when I've seen the short video on Reels
|
Questionnaire
|
CPI1
|
|
I will give recommendations regarding a
brand that I have seen in the short Reels video
|
Questionnaire
|
CPI2
|
|
I was about to make a purchase of the
product the first time I saw the brand in the Reels short video
|
Questionnaire
|
CPI3
|
|
I intend to find out more about the brand I will buy
|
Questionnaire
|
CPI4
|
Source: Author
Processed Data (2022)
Data analysis technique
The
data analysis technique used in this study is SEM (structural equation
modeling) with partial least squares (PLS). Partial least squares (PLS) is a
multivariate statistical technique that performs comparisons between multiple
dependent variables and multiple independent variables. together. (Hair, Hult, Ringle, & Sarstedt, 2016). The software used as a data processing
tool in this research is SmartPLS 3 software. The
test in this study consists of evaluating the outer model and inner model. The
outer model is used to display the relationship between latent variables and
indicators, while the inner model is used to display the relationship between
latent variables and latent variables.
RESULTS AND DISCUSSION
A. Research result
Evaluation in Smart PLS
consists of evaluation of the outer model (measurement model) and evaluation of
the inner model (structural model).
1.
Test the Measurement Model (Outer Model)
The measurement model is
a measurement to assess the validity and reliability of the model. Through the
algorithm iteration process, the measurement model parameters (convergent
validity, discriminant validity, composite reliability, and Cronbach's alpha)
were obtained, including the R2 value as a parameter for the accuracy of the
prediction model (Abdillah
et al., 2015). The results of the
measurement diagram (outer model) in this study can be seen in the figure, as
follows:
Figure 3.
Smart PLS 3.0 Output
Source: Author
Processed Data (2022)
The following
parameters are used in the measurement model test test
(outer model), as follows:
a) Convergent
Validity
Convergent validity, relates to the principle
that measurements of a construct should be highly correlated. Convergent
validity occurs when the scores obtained from two different instruments that
measure the same construct have a high correlation. The convergent validity
test in PLS with reflective indicators is considered fundamental on the loading
factor indicator that measures the construct, stating the rule of thumb used
for the outer loading convergent validity is > 0.7, and the AVE value which
is declared valid is 0.50 or higher, according to Hair figures. above 0.50 can
indicate that the construct explains at least 50% of the variance of each item.
Convergent validity all constructs in this study are valid (Hair
et al., 2019). Based on the results of data processing, the
convergent validity results obtained with the loading factor and AVE values, as
follows:
Table 2
Initial Loading Factor Value
|
Variable
|
Indicator
|
Factor
Loading
|
Conclusion
|
|
Interesting Content
|
IC1 ← IC
|
0.931
|
Valid
|
|
IC2 ← IC
|
0.703
|
Valid
|
|
IC3 ← IC
|
0.932
|
Valid
|
|
Scenario Based Experience
|
SBE1 ← SBE
|
0.802
|
Valid
|
|
SBE2 ← SBE
|
0.806
|
Valid
|
|
SBE3 ← SBE
|
0.791
|
Valid
|
|
SBE4 ← SBE
|
0.748
|
Valid
|
|
User Participation Interaction
|
UPI1 ← UPI
|
0.912
|
Valid
|
|
UPI2 ← UPI
|
0.748
|
Valid
|
|
UPI3 ← UPI
|
0.879
|
Valid
|
|
UPI4 ← UPI
|
0.884
|
Valid
|
|
Perceived Usefulness
|
PU1 ← PU
|
0.623
|
Invalid
|
|
PU2 ← PU
|
0.956
|
Valid
|
|
PU3 ← PU
|
0.498
|
Invalid
|
|
PU4 ← PU
|
0.589
|
Invalid
|
|
PU5 ← PU
|
0.630
|
Invalid
|
|
PU6 ← PU
|
0.953
|
Valid
|
|
PU7 ← PU
|
0.626
|
Invalid
|
|
PU8 ← PU
|
0.964
|
Valid
|
|
Perceived
Enjoyment
|
PE1 ← PE
|
0.820
|
Valid
|
|
PE2 ← PE
|
0.726
|
Valid
|
|
PE3 ← PE
|
0.782
|
Valid
|
|
PE4 ← PE
|
0.775
|
Valid
|
|
Involvement of Celebrity
|
IOC1← IOC
|
0.927
|
Valid
|
|
IOC2← IOC
|
0.706
|
Valid
|
|
IOC3← IOC
|
0.852
|
Valid
|
|
IOC4← IOC
|
0.894
|
Valid
|
|
IOC5← IOC
|
0.895
|
Valid
|
|
Consumer Brand Attitude
|
CBA1 ← CBA
|
0.852
|
Valid
|
|
CBA2 ← CBA
|
0.730
|
Valid
|
|
CBA3 ← CBA
|
0.898
|
Valid
|
|
CBA4 ← CBA
|
0.758
|
Valid
|
|
CBA5 ← CBA
|
0.835
|
Valid
|
|
Consumer Purchase Intention
|
CPI1 ← CPI
|
0.812
|
Valid
|
|
CPI2 ← CPI
|
0.825
|
Valid
|
|
CPI3 ← CPI
|
0.815
|
Valid
|
|
CPI4 ← CPI
|
0.763
|
Valid
|
Source:
Author Processed Data (2022)
Based
on the table above, it can be seen that there are several indicators that have
a loading factor value of <0.700, so it needs to be eliminated and retested
the convergent loading factor algorithm until the results are declared valid,
namely as follows:
Table 3
Final Loading Factor Value
|
Variable
|
Indicator
|
Factor Loading
|
Conclusion
|
|
Interesting Content
|
IC1
← IC
|
0.966
|
Valid
|
|
IC3
← IC
|
0.971
|
Valid
|
|
Scenario-based experience
|
SBE1
← SBE
|
0.803
|
Valid
|
|
SBE2
← SBE
|
0.805
|
Valid
|
|
SBE3
← SBE
|
0.788
|
Valid
|
|
SBE4
← SBE
|
0.750
|
Valid
|
|
User Participation
Interaction
|
UPI1
← UPI
|
0.955
|
Valid
|
|
UPI3
← UPI
|
0.855
|
Valid
|
|
UPI4
← UPI
|
0.929
|
Valid
|
|
Perceived Usefulness
|
PU2
← PU
|
0.991
|
Valid
|
|
PU6
← PU
|
0.986
|
Valid
|
|
PU8
← PU
|
0.996
|
Valid
|
|
Perceived Enjoyment
|
PE1
← PE
|
0.819
|
Valid
|
|
PE2
← PE
|
0.727
|
Valid
|
|
PE3
← PE
|
0.779
|
Valid
|
|
PE4
← PE
|
0.776
|
Valid
|
|
Involvement of Celebrity
|
IOC1←
IOC
|
0.973
|
Valid
|
|
IOC4←
IOC
|
0.936
|
Valid
|
|
IOC5←
IOC
|
0.944
|
Valid
|
|
Consumer Brand Attitude
|
CBA1
← CBA
|
0.842
|
Valid
|
|
CBA2
← CBA
|
0.740
|
Valid
|
|
CBA3
← CBA
|
0.890
|
Valid
|
|
CBA4
← CBA
|
0.768
|
Valid
|
|
CBA5
← CBA
|
0.837
|
Valid
|
|
Consumer Purchase
Intention
|
CPI1
← CPI
|
0.893
|
Valid
|
|
CPI1
← CPI
|
0.862
|
Valid
|
|
CPI3
← CPI
|
0.851
|
Valid
|
Source: Author Processed Data
(2022)
Based
on the results from the table above, the results of the evaluation of the final
results of convergent validity with a loading factor, by removing some
indicators that have a value smaller than 0.700 and re-estimating, it is
obtained that all indicators/items have a loading factor that has a value of
more than 0.7, so that it can be declared valid. In addition, convergent
validity can be measured by the average variance extracted (AVE), the AVE value
which is declared valid is > 0.5, a value > 0.50 can indicate that the
construct explains at least 50% of the variance of each item. Based on the
results of data processing, the results of the AVE value are obtained, as
follows:
Table 4
Average Variance Extracted
(AVE)
|
Variable
|
AVE
|
|
Interesting Content
|
0.938
|
|
Scenario-based experience
|
0.619
|
|
User Participation Interaction
|
0.835
|
|
Perceived Usefulness
|
0.982
|
|
Perceived Enjoyment
|
0.602
|
|
Involvement of Celebrity
|
0.905
|
|
Consumer Brand Attitude
|
0.667
|
|
Consumer Purchase Intention
|
0.755
|
Source: Author Processed
Data (2022)
Based on the table above,
the results of the calculation of convergent validity with AVE, obtained that
the AVE value of each variable has a value of more than 0.50. So it can be stated that the data in this study have met the
criteria of convergent validity.
b) Discriminant Validity
Discriminant validity is carried
out to assess the extent to which the construct is empirically different from
other constructs in the structural model. Related discriminant validity occurs when
two different instruments that measure two predicted uncorrelated constructs
result in an uncorrelated score. The criteria for testing discriminant validity
use the Heterotrait-Monotrait Ratio (HTMT) matrix in
PLS. Henseler et al. (2015) proposed to test discriminant
validity using HTMT with a value of less than 0.9. The HTMT value for each
variable in this study was below 0.9, meaning that the indicators were
appropriate for testing each construct. Based on the results of data
processing, the results obtained discriminant validity with the Heterotrait-Monotrait Ratio (HTMT) matrix method, as
follows:
Table 5
Heterotrait-Monotrait Ratio
(HTMT)
|
|
CBA
|
CPI
|
IC
|
IOC
|
PE
|
PU
|
SBE
|
UPI
|
|
Consumer Brand
Attitude
|
|
|
|
|
|
|
|
|
|
Consumer Purchase
Intention
|
0.779
|
|
|
|
|
|
|
|
|
Interesting Content
|
0.899
|
0.586
|
|
|
|
|
|
|
|
Involvement of
Celebrity
|
0.895
|
0.559
|
0.012
|
|
|
|
|
|
|
Perceived Enjoyment
|
0.896
|
0.881
|
0.721
|
0.695
|
|
|
|
|
|
Perceived Usefulness
|
0.831
|
0.455
|
0.510
|
0.497
|
0.677
|
|
|
|
|
Scenario Based
Experience
|
0.169
|
0.811
|
0.845
|
0.899
|
0.891
|
0.892
|
|
|
|
User Participation
Interaction
|
0.039
|
0.638
|
0.888
|
0.813
|
0.830
|
0.623
|
0.893
|
|
Source: Author Processed
Data (2022)
Based on
the table above, the results of the discriminant test with the Heterotrait-Monotrait Ratio (HTMT) matrix, the results show
that the correlation between variables has a value <0.9, therefore, the
variables in this study can be declared to meet the criteria of discriminant
validity.
c) Construct Reliability
The
construct reliability test was measured in two ways, namely composite
reliability and cronbach alpha. According to Hair et al. (2014), Cronbach
alpha which has a value > 0.70 is said to have a reliable construct.
Meanwhile, higher composite reliability indicates that the level of reliability
is higher. Reliable value on composite reliability > 0.70. Based on the
results of data processing, the results obtained construct reliability, as
follows.
Table 6
Construct Reliability
|
Variabel
|
Composite Reliability
|
Cronbach Alpha
|
|
Interesting Content
|
0.968
|
0.934
|
|
Scenario-based experience
|
0.867
|
0.795
|
|
User Participation Interaction
|
0.938
|
0.900
|
|
Perceived Usefulness
|
0.994
|
0.991
|
|
Perceived Enjoyment
|
0.858
|
0.782
|
|
Involvement of Celebrity
|
0.966
|
0.947
|
|
Consumer Brand Attitude
|
0.909
|
0.874
|
|
Consumer Purchase Intention
|
0.902
|
0.839
|
Source:
Author Processed Data (2022)
Based
on the table above, the results of the reliability test show that all variables
in this study have a value > 0.7 which means that the variables used are
reliable. The variable with the highest composite reliability value and
Cronbach alpha is in the perceived usefulness variable with a CR value of 0.994
and a CA of 0.991, while the lowest value is the perceived enjoyment variable,
with a CR value of 0.858 and a CA of 0.782.
d) Multicollinearity
(VIF)
Inner
VIF Value is a test to find out whether between indicators has
multicollinearity. According to Hair
et al. (2014),
VIF which has a value lower than 5 indicates that the indicator does not
experience multicollinearity. This study has between indicators in this study
do not experience multicollinearity. Based on the results of data processing,
the results of the multicollinearity test are obtained, as follows:
Table 7
Collinearity (VIF)
|
VIF
|
|
CBA1
|
4.331
|
|
CBA2
|
1.666
|
|
CBA3
|
4.805
|
|
CBA4
|
2.128
|
|
CBA5
|
2.621
|
|
CPI1
|
2.612
|
|
CPI2
|
2.336
|
|
CPI3
|
1.638
|
|
IC1
|
4.302
|
|
IC3
|
4.302
|
|
IOC1
|
4,984
|
|
IOC4
|
4.300
|
|
IOC5
|
4.139
|
|
PE1
|
1.911
|
|
PE2
|
1.565
|
|
PE3
|
1.674
|
|
PE4
|
1.431
|
|
PU2
|
4.923
|
|
PU6
|
4.782
|
|
PU8
|
4.608
|
|
SBE1
|
1.778
|
|
SBE2
|
1.742
|
|
SBE3
|
1.627
|
|
SBE4
|
1.551
|
|
UPI1
|
4.006
|
|
UPI3
|
1.951
|
|
UPI4
|
4.059
|
Source:
Author Processed Data (2022)
Based
on the table above, it can be seen that each variable indicator has a VIF value
<5, so it can be stated that there is no correlation between the data, it
can be declared free of multicollinearity.
2. Structural Model Test (Inner
Model)
The structural model in PLS is evaluated using R2
for the dependent construct, the path coefficient value or the t-value of each
path to test the significance between constructs in the structural model, the
next step is to evaluate the structural model to see the significance of the
relationship between constructs/variables. This can be seen from the path
coefficient which describes the strength of the relationship between
constructs. The sign or direction on the path (path coefficient) must be in
accordance with the hypothesized theory, its significance can be seen in the t
test or CR (critical ratio) obtained from the bootstrap process (resampling
method). The structural model (inner model) is carried out by testing the
R-square, Q-square, and path coefficients using the SmartPLS
software. The path diagram of the inner model in this study can be seen in the
figure, as follows:
Figure 4. Inner Model
Source:
Author Processed Data (2022)
The
following are the parameters used in the structural model test test (inner model), as follows:
a) R-Square (R2)
R-Square, measuring explanatory power and predictive
accuracy on research constructs can be done using the R-Square test. Hair
et al. (2014) say
that the R-square value of 0.75 has a strong influence, a value of 0.5
indicates a moderate effect, and a value of 0.25 indicates a weak effect. Based
on the results of data processing, the results of the r-square are obtained, as
follows:
Table 8
R-Square
(R2) values
|
Variable
|
R-Square
|
|
Consumer Brand Attitude
|
0.985
|
|
Consumer Purchase Intention
|
0.459
|
Source:
Author Processed Data (2022)
Based
on the table above, it shows that the rsquare value
on the intervening variable consumer brand attitude is 0.985, which indicates
that it is in the strong category, this indicates that consumer brand attitude
can be explained by 98.5% by variables interesting content, scenario-based
experience, user participation interaction, perceived usefulness, perceived
enjoyment and involvement of celebrity, while the remaining 1.5% can be
explained by other variables outside the research model. Meanwhile, the rsquare value of the dependent variable consumer purchase
intention is 0.459, which indicates that it is in the moderate/moderate
category, this indicates that consumer purchase intention can be explained by
45.9% by the consumer brand attitude variable, while the remaining 54.1% can be
explained by the variables other variables outside the
research model
b) Q-Square
The Q-square
value is used to show predictive relevance. According to Hair et al., (2019), the Q-square value which has a value range of 0 has a small
meaning, 0.25 medium and 0.5 large. Meanwhile, a large Q-square value > 0.5
indicates good predictive relevance. The results of the calculation of the
Q-Square value are as follows:
�Q-Square�� =��� 1 � (1 � R21) x (1 � R22)
���
���������������=��� 1 � (1 � 0.985) x (1
� 0.459)
����������������� �=��� 1 �� 0.008
������������������ =���� 0.992 atau 99.2%
Based
on the results of these calculations, the obtained Q-square results of 0.992 or
99.2%, so it can be stated that the variance of the magnitude of the diversity
of the research data used has a large predictive relevance, where changes in
the sample will not be affected.
c) Hypothesis Testing
The hypothesis in this study can
be seen from the calculation of the model using the PLS bootstrapping
technique. Based on the data processing that has been done, these results can
be used in answering the hypothesis in this study. Hypothesis testing is done
by looking at the t-Statistics value and the P-Values value. The
research hypothesis can be declared accepted if the direction of the path
coefficient shows results that are in accordance with the initial hypothesis
and t-statistics > t table (1,984), Meanwhile, the smaller the P-Values, the
stronger the evidence that the null hypothesis must be rejected. P-Values
that show a number < 0.05 are considered statistically
significant. The following are the results of hypothesis testing obtained in
this study through the inner model, consisting of the direct effect and
indirect effect hypotheses.
Table 9
Direct Effect Hypothesis Test Results
|
Hypothesis
|
Standard Path Coefficient
|
T-Statistics
|
P-Value
|
Significance Description
|
Results
|
|
|
H1
|
Interesting Content� -> Consumer Brand Attitude
|
0.073
|
1.138
|
0.128
|
Not significant
|
Hypothesis Not Supported
|
|
H2
|
Scenario Based Experience� -> Consumer Brand Attitude
|
0.455
|
7.061
|
0.000
|
Significant
|
Hypothesis Not Supported
|
|
H3
|
User Participation Interaction -> Consumer
Brand Attitude
|
0.165
|
2.706
|
0.004
|
Significant
|
Hypothesis Supported
|
|
H4
|
Perceived Usefulness -> Consumer Brand Attitude
|
0.125
|
3.433
|
0.000
|
Significant
|
Hypothesis Supported
|
|
H5
|
Perceived Enjoyment -> Consumer Brand Attitude
|
0.102
|
4.152
|
0.000
|
Significant
|
Hypothesis Supported
|
|
H6
|
Involvement of Celebrity -> Consumer Brand
Attitude
|
0.166
|
3.017
|
0.001
|
Significant
|
Hypothesis Supported
|
|
H7
|
Consumer Brand Attitude -> Consumer Purchase
Intention
|
0.678
|
12.952
|
0.000
|
Significant
|
Hypothesis
Supported
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Source:
Author Processed Data (2022)
Based
on the results of testing the direct effect hypothesis, it is concluded that
the hypothesis testing between variables is as follows:
1) Hypothesis
Testing H1: Effect of Interesting Content on Consumer Brand Attitude
The
results of hypothesis testing show that the t-statistics value obtained is
1.138, where the t-statistics value < t-table value is 1.984
(1.138<1.984) and the significance is 0.128>0.05. Thus, it can be stated
that interesting content has no significant positive effect on consumer brand
attitude.
2) Hypothesis
Testing H2: Effect of Scenario Based Experience on Consumer Brand Attitude
The
results of hypothesis testing show that the t-statistics value obtained is
7.061, where the t-statistics value > t-table value is 1.984
(7.061>1.984) and the significance is 0.000<0.05. Thus, it can be stated
that scenario-based experience has a significant positive effect on consumer
brand attitudes.
3) Hypothesis
Testing H3: The Effect of User Participation Interaction on Consumer Brand
Attitude
The
results of hypothesis testing show that the t-statistics value obtained is
2.706, where the t-statistics value > t-table value is 1.984
(2.706>1.984) and the significance is 0.004<0.05. Thus, it can be stated
that user participation interaction has a significant positive effect on
consumer brand attitudes.
4) Hypothesis
Testing H4: The Effect of Perceived Usefulness on Consumer Brand Attitude
The
results of hypothesis testing show that the t-statistics value obtained is
3.433, where the t-statistics value > t-table value is 1.984
(3.433>1.984) and the significance is 0.000<0.05. Thus, it can be stated
that perceived usefulness has a significant positive effect on consumer brand
attitude.
5) Hypothesis
Testing H5: The Effect of Perceived Enjoyment on Consumer Brand Attitude
The
results of hypothesis testing show that the t-statistics value obtained is
4.152, where the t-statistics value > t-table value is 1.984
(4.152>1.984) and the significance is 0.000<0.05. Thus, it can be stated
that perceived enjoyment has a significant positive effect on consumer brand
attitude.
6) Hypothesis
Testing H6: The Influence of Involvement of Celebrity on Consumer Brand
Attitude
The
results of hypothesis testing show that the t-statistics value obtained is 3.017,
where the t-statistics value > t-table value is 1.984 (3.017>1.984) and
the significance is 0.001<0.05. Thus, it can be stated that the involvement
of celebrity has a significant positive effect on consumer brand attitude.
7) Hypothesis
Testing H7: The Effect of Consumer Brand Attitude on Consumer Purchase
Intention
The
results of hypothesis testing show that the t-statistics value obtained is
12.952, where the t-statistics value > t-table value is 1.984
(12.952>1.984) and the significance is 0.000<0.05. Thus, it can be stated
that consumer brand attitude has a significant positive effect on consumer
purchase intention.
Furthermore, testing the indirect effect
hypothesis, the influence of the independent variable on the dependent variable
through the intervening variable, the results obtained are as follows:
Table 10
Indirect Effect
Hypothesis Test Results
|
Hypothesis
|
Standard Path Coefficient
|
T-Statistics
|
P-Value
|
Significance Description
|
Results
|
|
H1
|
Interesting Content� -> Consumer Brand Attitude ->
Consumer Purchase Intention
|
0.049
|
1.130
|
0.130
|
Not significant
|
Hypothesis Not Supported
|
|
H2
|
Scenario Based Experience� -> Consumer Brand Attitude ->
Consumer Purchase Intention
|
0.309
|
6.587
|
0.000
|
Significant
|
Hypothesis Supported
|
|
H3
|
User Participation Interaction -> Consumer
Brand Attitude -> Consumer Purchase Intention
|
0.112
|
2.743
|
0.003
|
Significant
|
Hypothesis Supported
|
|
H4
|
Perceived Usefulness -> Consumer Brand Attitude
-> Consumer Purchase Intention
|
0.085
|
3.062
|
0.001
|
Significant
|
Hypothesis Supported
|
|
H5
|
Perceived Enjoyment -> Consumer Brand Attitude
-> Consumer Purchase Intention
|
0.069
|
3.835
|
0.000
|
Significant
|
Hypothesis Supported
|
|
H6
|
Involvement of Celebrity -> Consumer Brand
Attitude -> Consumer Purchase Intention
|
0.113
|
2.862
|
0.002
|
Significant
|
Hypothesis Supported
|
Source:
Author Processed Data (2022)
Based
on the results of testing the indirect effect hypothesis, it is concluded that
the hypothesis testing between variables is as follows:
1)
There is no significant
positive effect of interesting content on consumer purchase intention through
consumer brand attitude.
2)
There is a significant
positive effect of scenario based experience on
consumer purchase intention through consumer brand attitude.
3)
There is a significant
positive effect of user participation interaction on consumer purchase
intention through consumer brand attitude.
4)
There is a significant
positive effect of perceived usefulness on consumer purchase intention through
consumer brand attitude.
5)
There is a significant
positive effect of perceived enjoyment on consumer purchase intention through
consumer brand attitude.
6)
There is a significant
positive effect of involvement of celebrity on consumer purchase intention
through consumer brand attitude.
CONCLUSION
The results of
hypothesis testing show that the independent scenario based
variable, user participation, perceived usefulness, perceived enjoyment,
involvement of celebrity directly (direct effect) has a positive and
significant effect on consumer brand attitude, and consumer brand attitude also
has a positive and significant influence on consumer purchase. intention.
Meanwhile, interesting content directly (direct effect) has no positive and
significant effect on consumer brand attitude.
The highest
coefficient value is shown in the effect of scenario based
experience on consumer brand attitude, with a path coefficient value of 0.455,
and the lowest value is found in the effect of interesting content on consumer
brand attitude with a path coefficient value of 0.073. While the influence of
the intervening variable consumer brand attitude on the dependent variable
consumer purchase intention obtains a path coefficient value of 0.678.
For the results of
testing the indirect effect hypothesis, the effect of the independent variable
on the dependent variable through the intervening variable, which obtained the
highest value was in the influence of scenario based
experience on consumer purchase intention through consumer brand attitude with
a path coefficient value of 0.309.
REFERENCES
Abdillah, Willy, Hartono, & Jogiyanto. (2015). Partial Least Suare
(PLS) - Alternatif Structural equation Modeling (SEM) dalam penelitian Bisnis.
Bandung: Andi Offset.
Baskara, I. M. A., & Sukaatmadja, I. P. G. (2016). Pengaruh online
trust dan perceived enjoyment terhadap online shopping satisfaction dan
repurchase intention Lazada Indonesia. Udayana University. Google Scholar
Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user
acceptance of information technology. MIS Quarterly, 319�340. Google Scholar
Ha, N. M., & Lam, N. H. (2017). The effects of celebrity endorsement
on customer�s attitude toward brand and purchase intention. International
Journal of Economics and Finance, 9(1), 64�77. Google Scholar
Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2014). Multivariate
Data Analysis. Pearson Education Limited. Google Scholar
Hair, J. F., Risher, J. J., Sarstedt, M., & Ringle, C. M. (2019). When
to use and how to report the results of PLS-SEM. European Business Review,
31(1), 2�24. Google Scholar
Henseler, J., Ringle, C. M., & Sarstedt, M. (2015). A new criterion
for assessing discriminant validity in variance-based structural equation
modeling. Journal of the Academy of Marketing Science, 43(1),
115�135. Google Scholar
Kriyantono, R. (2020). Teknik praktis riset komunikasi kualitatif dan
kuantitatif. Jakarta: Prenada Media.
Liu, G., Gao, P., Li, Y., & Zhang, Z. (2019). Research on the
influence of social media short video marketing on consumer brand attitude. 2019
5th International Conference on Social Science and Higher Education (ICSSHE
2019), 433�438. Google Scholar
Maksimova, A., & Savolainen, A. (2021). Instagram as a Digital
Marketing tool for fashion brands. Google Scholar
Sharp, G., & Gerrard, Y. (2022). The body image �problem� on social
media: Novel directions for the field. Body Image, 41, 267�271. Elsevier
Sugiyono. (2018). Metode Penelitian Kuantitatif, Kualitatif, dan
R&D. Bnadung: Alfabeta. Google Scholar
Winarso, W. (2020). Technology, social media and behaviour of young
generation in Indonesia; a conseptual paper. International Journal of
Scientific & Technology Research, 9(4), 986�989. Google Scholar
Wuisan, D., Kananda, I. L., Saputra, M. A., & Dinata, M. R. (2020).
The Effect of Celebrity Endorsement on Purchase Intention of Adidas Products
Mediated by Customer�s Attitude Toward Brand. Conference Series, 3(1),
525�533. Google Scholar
Zuharah, L. A. Z. L. A., & Tunggali, A. A. P. P. W. (2022). M Motives
For Using Reels Instagram Social Media Among Children. COMMICAST, 3(3),
259�265. Google Scholar
IMPACT OF REELS VIDEO MARKETING ON
CUSTOMERS� PURCHASE INTENTION