Service Quality for Digital Wallet in Indonesia Using Sentiment Analysis and Topic Modelling
Winda Aulia Deviani1*, Krishna Kusumahadi1, Eva Nurhazizah1
1 Faculty of Economics and Business, Telkom University, Bandung, Indonesia
*Corresponding Author: [email protected]
Accepted: 15 March 2022 | Published: 1 April 2022
DOI:https://doi.org/10.55057/ijbtm.2022.4.1.6
__________________________________________________________________________________________
Abstract: With the digital wallet, it is able to make it easy for users to store, send, and pay directly through the application. The Fintech Report by DSResearch states that Dana and ShopeePay are digital wallet products that have managed to occupy the top position in terms of daily usage frequency. However, the quality of Dana and ShopeePay services is considered less than optimal because many Dana and ShopeePay users have tweeted their complaints via Twitter social media. So that Dana and ShopeePay need to maintain service quality to maintain the loyalty of their users. This study aims to determine user sentiment towards the quality of Dana and ShopeePay services based on the e-servqual dimension and to find out what topics are formed in each e-servqual dimension to measure the service quality of Dana and ShopeePay. The data sources for this research are Dana and ShopeePay user-generated content delivered through social media Twitter. The data retrieval technique is crawling user tweets containing the keywords "@danawallet" and "@ShopeePay_ID" with a time span of 23 October 2021 to 29 December 2021. The data obtained will be classified based on the e- servqual dimension using Naive Bayes and the dataset will be analyzed using sentiment analysis and topic modeling. The results in this study show that negative sentiments dominate Dana and ShopeePay's e-servquals on the dimensions of Efficiency, System Availability, Fulfillment, and Privacy, as well as topics and words that have a negative connotation on the services provided by Dana and ShopeePay.The results of this research can be used by Dana and ShopeePay as an evaluation of service quality, especially on the e-servqual dimension to increase user satisfaction to maintain user loyalty and improve user perceptions.
Keywords: digital wallet, e-service quality, sentiment analysis, topic modeling
___________________________________________________________________________
1. Introduction
Ease of accessing the internet supports the convenience of the public in non-cash payments.
Non-cash payments are an option for the public to make payments for transactions for both goods and services(Foster, 2020). Non-cash payments can be in the form of a digital wallet, a server-based digital wallet service to accommodate a certain amount of non-cash money from users, allowing users to store, send, and pay directly through the application (DSResearch, 2020). Some digital wallet companies include Dana and ShopeePay. Beating the position of GoPay, OVO, and LinkAja in terms of daily usage frequency, ShopeePay and Dana managed to occupy the first and second positions as products that are often used every day (DSResearch, 2020). This creates quite a competitive competition between digital wallet services. Based on the statement, digital wallet companies can implement strategies to maintain customer loyalty.
Research by (Foster, 2020) states that improving the quality of e-wallet services to consumers can increase customer loyalty. If users experience problems when using a digital wallet, customer service must be able to assist users in overcoming these obstacles (Foster, 2020).
Therefore, an analysis of the quality of Dana and ShopeePay services is needed to find out whether the digital wallet service is in line with expectations or not.
User perception is very important to determine the strategy in terms of service quality in order to maintain the loyalty of the users. According to (Mothersbaugh & Hawkins, 2016)Perception is a process that begins with the exposure and attention of users to marketing stimuli and ends with the user's interpretation of these marketing stimuli. The strategy that can be used to analyze service quality is using the e-servqual method. According to (Parasuraman et al., 2005),the method used to measure service quality is e-servqual. E-servqual is the ability of a site to provide effective and efficient facilities for online shopping, online purchases, and in the acquisition of goods or services (Parasuraman et al., 2005). There are several dimensions of e-servqual proposed by (Parasuraman et al., 2005) namely efficiency or efficiency, fulfillment or fulfillment, system availability or system availability, and privacy or privacy. So that in order to improve the quality of digital wallet services, companies must be able to meet the dimensions of e-servqual.
Digital wallet companies have different ways of communicating with their users. Companies can use social media to communicate with their users. One of the social media used by Dana and ShopeePay is Twitter. Dana and ShopeePay users can tell about the problems they experience when using a digital wallet, ask for information or even provide an assessment regarding the digital wallet via tweets by mentioning Dana and ShopeePay usernames, namely
@danawallet and @ShopeePay_ID. However, based on the tweets of Dana and ShopeePay users, the quality of the services provided is considered less than optimal. Research(Ogi et al., 2021) states that Dana's digital wallet has a negative sentiment percentage of 82.22% and t.
Thets of wordcloud data processing conclude that many Dana users are disappointed with customer service in hanng user complaints. Research (Meidita et al., 2018)states that users are less satisfied with ShopeePay's payment features. Based on the pre-research conducted by the researcher, Dana has a negative sentiment percentage of 56% and ShopeePay has a negative sentiment percentage of 81%. Topics generated by Dana and ShopeePay are dominantly negative. Based on this, the quality of Dana and ShopeePay services is considered less than optimal. So this study applies the e-servqual method to find out how user sentiment is on the quality of Dana and ShopeePay services based on the e-servqual dimensions and what topics are formed on each e-servqual dimension to measure the quality of Dana and ShopeePay services.
2. Literature Review
2.1 Perception
Perception can receive stimuli related to the surrounding environment and the individual's circumstances(Rizal, 2020).Companies can communicate by utilizing the internet, product placement in films, and mobile phone screen displays (Malhotra, 2018). Based on the book Consumer Behavior Bulding Marketing Strategy Thirteenth Edition by (Mothersbaugh &
Hawkins, 2016)perception is a process that begins with the exposure and attention of consumers to marketing stimuli and ends with the consumer's interpretation of these marketing stimuli. Everyone's perception can be different even though the object is the same because there are three processes of understanding: exposure, attention, and interpretation.
2.2 E-Service Quality
According to (Parasuraman et al., 2005)e-service quality is an electronic-based service that is used to facilitate shopping, purchasing and delivery of products and services effectively and efficiently. According to Wu (2014) in (Magdalena & Jaolis, 2018) e-servqual is a service provided on the internet network to facilitate shopping, purchasing, and distribution activities.
The e-servqual model is the most developed online service quality model(Nurapipah, 2021).
Therefore, online customers expect the level of service quality to be better than traditional customers (Irwansyah & Mappadeceng, 2018). To measure the quality of electronic services, four main dimensions include efficiency, fulfillment, system availability, and privacy. These dimensions are relevant and comprehensively meet the need to evaluate the quality of electronic services (Billyarta & Sudarusman, 2021).
According to (Parasuraman et al., 2005)efficiency is the ability of customers to access a website, search for the desired product and information related to that product, and leave the site with minimal effort. When customers experience confusion during the search process, the customer will stop (Magdalena & Jaolis, 2018). The second dimension is fulfillment.
According to (Billyarta & Sudarusman, 2021) fulfillment is the accuracy of service promises, product inventory availability, and product delivery according to schedule. According to (Magdalena & Jaolis, 2018) fulfillment is the ability of a website to correct errors that occur during transactions. In fulfillment, the company must provide actual performance with what was promised through the website (Parasuraman et al., 2005).
The third dimension is system availability. According to (Parasuraman et al., 2005) system availability is the ability of a website to function as it should. The features provided by the system must be able to avoid system failure (Nurapipah, 2021). The fourth dimension is privacy. According to (Billyarta & Sudarusman, 2021) privacy is the level of website security and protection of customer information. Credit or debit card information must also be safe and not leaked to outside parties (Irwansyah & Mappadeceng, 2018). When the services provided to customers are done well, it can build trust and confidence in customers (Magdalena & Jaolis, 2018).
2.3 Big Data
Based on the book Artificial Intelligence Basics by (Taulli, 2019)big data is a technology category that involves processing large amounts of data. Big data represents complex data sets that require large storage, processing time, and scalable units to store and process (Hassanien
& Darwish, 2021). Using big data can make the right decisions by analyzing data that influence each other (Alani et al., 2021).To analyze the data, there are several stages, namely obtaining data sources, pre-processing data, and building models. At the stage of building the model, you can choose machine learning. Machine learning is the study of algorithms that learn from data to make decisions (Alani et al., 2021). Machine learning allows machines to learn without being programmed (Akerar, 2019). Another benefit when using data analytics is to help marketing in segmentation or marketing mix (Akerar, 2019). Using big data can also analyze data sourced from social media to gain customer insights (Masrury et al., 2019). In research (Alamsyah &
Bernatapi, 2019)using big data to help companies develop customer experience management.
2.4 User Generated Content
In research (Masrury et al., 2019)user generated content is content created by companies to promote products published on public sites or social media that can be accessed by a group of people. UGC can be in the form of text, photos, videos, and microblog uploads on social media (Pinuji, 2019). According to Manap & Adzharudin (2013) in (Bahtar & Muda, 2016)UGC is
also known as electronic word of mouth. UGC comes from users who voluntarily provide useful information or data (Amalia & Sudiwijaya, 2020). According to the Organization for Economic Co-Operation and Development (2007) in (Rubyanti & Irwansyah, 2020)there are characteristics in UGC, namely that it can be uploaded and can be accessed by the public, enter values from users or uploaders creatively, and is a form of expression.
UGC can also be called feedback because when previous buyers share their experiences using a product or service and many people see it, it allows for potential buyers (Bahtar & Muda, 2016). The role of the user is not only as a consumer but also as a content creator (Rubyanti &
Irwansyah, 2020). Conversations conveyed by users through social media are also referred to as UGC indicators (Amalia & Sudiwijaya, 2020). According to Barmhart, B (2019) in research (Ogi et al., 2021)UGC delivered by ,users via Twitter is easier to understand, in contrast to o thana because users are asked to explain something quickly, briefly, and concisely. According to Estrella-Ramon and Ellis Chadwick (2017) in (Pinuji, 2019)there are two types of UGC, namely positive and negative where a positive UGC is a favorable experience and recommendation to buy certain products while a negative UGC is an unfavorable experience and recommendation for do not buy certain products (Pinuji, 2019).
3. Methodology
3.1 Population and Sample
The population used in this study is Twitter users who tweet with the keywords "@danawallet"
and "@ShopeePay_ID". The sample used in this study is tweets containing the keywords
"@danawallet" and "@ShopeePay_ID" which are relevant to the e-servqual dimension uploaded in the period from October 23, 2021 to December 29, 2021.
3.2 Data Collection
The data used in this study is secondary data. The data taken are tweets containing the keywords
"@danawallet" and "@ShopeePay" which are relevant to the e-servqual dimension uploaded in the period from October 23, 2021 to December 29, 2021. The data is taken by crawling technique on Twitter social media using Application Programming Twitter and Google Colab interfaces.
3.3 Data Analysis Techniques 3.3.1 Data Pre-Processing
The data that has been collected still needs to be done in the data pre-processing stage so that it can perform data analysis. Columns in the dataset used are only text data. Applications that will be used RapidMiner Studio 9.1. At the pre-processing stage, several activities are carried out, including:
a. Transform Cases: This stage changes the letters in the tweet to lowercase.
b. Tokenization: This stage is done to separate sentences into separate words or phrases.
c. Stopword Removal: At this stage, filtering is carried out to remove words that have no meaning. Researchers will use documents containing descriptive words such as which, namely, are, and so on to produce meaningful words in the dataset.
d. Stemming: This stage changes the words into standard words in Indonesian.
3.3.2 Classification Based On The Dimensions Of E-servqual
Data that has gone through the pre-processing stage will be classified based on e-servqual. To classify the data, machine learning is needed. The process in machine learning is in the form of training data and testing data. However, to perform data training and data testing, a labeling
process is needed to determine the data based on the characteristics of the data.The results to be obtained are datasets that have been classified based on the dimensions of e-servqual.The following is a labeling process based on the dimensions of e-servqual which can be seen in Table 2 below.
Table 2: Labelling Process Based on The Dimensions of E-servqual
Text Dimensions of E-servqual
Tolong cek DM ya admin Efficiency Senang banget kirim uang bebas biaya
admin
Fulfillment Kenapa transaksi pembayaran saya
pending
System Availability Keamanan dan kerahasiaan nomor 1 Privacy
3.3.3 Sentiment Analysis
According to (Masrury et al., 2019)Sentiment analysis is used to see the emotions of users related to applications and can detect the context of the patterns contained in the text. In analyzing user sentiment, machine learning algorithms can be used because they are considered to have high accuracy in classifying text. In this study, the algorithm used is Naïve Bayes. The technique in this sentiment analysis is to classify the text into two classifications, namely positive and negative. The purpose of using sentiment analysis is to determine user sentiment on the quality of Dana and ShopeePay services based on the e-servqual dimension. The software used by researchers for data processing is RapidMiner Studio 9.1. Datasets that have been classified based on the dimensions of e-servqual will be labeled with positive and negative sentiments. The following is the process of labeling training data which can be seen in Table 3 below.
Table 3: Labelling Process of Sentiment Analysis
Text Sentimen
Senang banget kirim uang bebas biaya admin
Positif Kenapa transaksi pembayaran saya
pending
Negatif
3.3.4 Topic Modelling
Based on research by (Masrury et al., 2019)Topic modeling or topic modeling aims to see the main topics submitted by users related to applications. With this topic modeling, developers can determine areas that need to be improved based on topics that users frequently discuss.
Topic Modeling is a way to let the text do the talking because the topics identified do not depend on the user's perspective or experience. The algorithm that can be used to perform topic modeling is Latent Dirichlet Allocation (LDA) which has a significant impact on Natural Language Processing and Machine Learning. In this research, topic modeling is used to find out what topics are formed in each e-servqual dimension to measure the quality of Dana and ShopeePay services.
The software that the researcher uses for data processing is Google Colab. The dataset that will be used is data that has been classified based on the dimensions of e-servqual. The dataset that will be used by researchers will be uploaded via Google Drive as a step to import data on Google Colab. The results of topic modeling are in the form of words that represent a particular topic. Words that will represent a topic will be chosen based on words that appear frequently.
In this study, the researcher chose thirty words that would represent the topic on each e-servqual dimension.
4. Discussion and Conclusion
4.1. Data Characteristics
The data collection process takes place from October 23, 2021 to December 29, 2021, sourced from crawling via Twitter social media and in the form of tweets in csv format using Google Colab. The data obtained are 7,929 tweets for Dana while 6,279 tweets for ShopeePay. The total tweets obtained were 14,208 tweets.
4.2. Data Pre-processing
The pre-processing stage is the stage to prepare raw data into data that is ready to be processed.
The results of pre-processing can be seen in Table 4below.
Table 4: Pre-processing Data Result
Digital Wallet Before Pre-processing After Pre-processing
Dana 7.929 data 4.454 data
ShopeePay 6.279 data 742 data
4.3. Classification Based On The Dimensions Of E-servqual
The results of the e-servqual dimension classification in Dana and ShopeePay can be seen at Table 5 below.
Table 5: Results of E-servqual Dimension Classification in Dana and ShopeePay Digital
Wallet
Dimensi E-servqual Jumlah
Efficiency Fulfillment System Availability
Privacy
Dana 1573 1080 1440 361 4454
ShopeePay 159 73 455 55 742
Based on the results of thdata classification resultst dimension in Dana is Efficiency. While in ShopeePay, the dominant dimension is System Availability.
4.4. Sentiment Analysis
Sentiment Analysis results on each Digital Wallet can be seen in Table 6 and Table 7 below.
Table 6: Result of Sentiment Analysis on Dana
Sentiment The Dimensions of E-servqual Total Efficiency Fulfillment System
Availability
Privacy
Positive 73 231 78 126 508
Negative 1500 849 1362 235 3946
Table 6 shows that the fulfillment dimension is dominated by positive sentiment, which is 45%
or 231 data from 508 data, while the efficiency dimension is dominated by negative sentiment, which is 37% or 1500 data from 3946. While on ShopeePay can be seen in Table 7.
Table 7: Result of Sentiment Analysis on ShopeePay
Sentiment The Dimensions of E-servqual Total
Efficiency Fulfillment System Availability
Privacy
Positive 3 27 1 1 32
Negative 156 46 454 54 710
Table 7 shows that the fulfillment dimension is dominated by positive sentiment, which is 85%
or 27 data from 32 data, while the system availability dimension is dominated by negative sentiment, which is 64% or 454 data from 710.
4.5. Topic Modelling
To see what topics appear based on dimension of e-servqual, can be seen in Figure 1, Figure 2, Figure 3, Figure 4, Figure 5, Figure 6, Figure 7, and Figure 8.
Figure 1: Main Topic of Dana Based on The Efficiency Dimension
Based on the words that often appear above, it can be said that the main topic of Dana based on the efficiency dimension shows that customer service is slow to respond to user complaints. Customer service that is slow to respond refers to the words 'tolong' and 'cek' when the user has submitted a complaint which refers to the words 'lapor' and 'udah'.
To see the main topic of Dana based on the system availability dimension, it can be seen in Figure 2 below.
Figure 2: Main Topic of Dana Based on The System Availability Dimension
Based on the words that often appear above, it can be said that the main topic of Dana based on the system availability dimension shows user failures in transactions, both top up balances, transfers, and buying credit. User failure refers to the word 'error'.
To see the main topic of Dana based on the fulfillment dimension, it can be seen in Figure 3 below.
Figure 3: Main Topic of Dana Based on The Fulfillment Dimension
Based on the words that often appear above, it can be said that the main topic of Dana based on the fulfillment dimension shows user dissatisfaction when using Dana. User dissatisfaction refers to the words 'drama' and 'banget'.
To see the main topic of Dana based on the privacy dimension, it can be seen in Figure 4 below.
Figure 4: Main Topic of Dana Based on The Privacy Dimension
Based on the words that often appear above, it can be said that the main topic of the Fund based on the privacy dimension shows that transactions or data of Dana users are not safe. Unsafe users refer to the word 'gimana'.
The main topic of ShopeePay based on the efficiency dimension can be seen in Figure 5 below.
Figure 5: Main Topic of ShopeePay Based on The Efficiency Dimension
These words represent the slow response of ShopeePay customer service in responding to ShopeePay user complaints. Words that represent the slow response of ShopeePay's customer service are 'sampe' and 'jam'.
The main topic of ShopeePay based on the system availability dimension can be seen in Figure 6 below.
Figure 6: Main Topic of ShopeePay Based on The System Availability Dimension
Based on the words that often appear above, it can be said that the main topic of ShopeePay based on the system availability dimension shows user failures in transactions and balance top ups. User failure refers to the words 'error' and 'gabisa'.
To see the main topic of ShopeePay based on the fulfillment dimension, it can be seen in Figure 7 below.
Figure 7: Main Topic of ShopeePay Based on The Fulfillment Dimension
Based on the words that often appear above, it can be said that the main topic of ShopeePay based on the fulfillment dimension shows user dissatisfaction when using ShopeePay. User dissatisfaction refers to the words 'deh', 'suka', and 'kesel'.
To see the main topic of ShopeePay based on the privacy dimension, it can be seen in Figure 8 below.
Figure 8: Main Topic of ShopeePay Based on The Privacy Dimension
Based on the words that often appear above, it can be said that the main topic of ShopeePay based on the privacy dimension shows that it is not safe when logging in. Insecure when logging in refers to the words 'cek', 'dm', and 'login'.
4.6. Conclusion
1) Tweets submitted by users contain positive and negative sentiments for each digital wallet.
Dana has 12% positive sentiment and 88% negative sentiment based on user tweets.
Meanwhile ShopeePay has 5% positive sentiment and 95% negative sentiment. Based on the e-servqual dimension, Dana has 75 positive sentiment data and 1500 negative sentiment data on the Efficiency dimension, 78 positive sentiment data and 1362 negative sentiment data on the System Availability dimension, 231 positive sentiment data and 849 negative sentiment data on the Fulfillment dimension, and 126 positive sentiment data and 235 negative sentiment data on the Privacy dimension. Meanwhile ShopeePay has 1 positive sentiment data and 454 negative sentiment data on the System Availability dimension, 156 positive sentiment data and 3 negative sentiment data on the Efficiency dimension, 27 positive sentiment data and 46 negative sentiment data on the Fulfillment dimension, and 1 positive sentiment data and 54 negative sentiment data on the Privacy dimension. From the results of the sentiment in the two digital wallets, it can be said that the services provided by Dana and ShopeePay are still not good.
2) Based on the topic modeling carried out, the main topics in each e-servqual dimension were obtained for each digital wallet. The Efficiency dimension Dana and ShopeePay have the same main topic, namely that digital wallets are considered unable to provide a fast and maximum response in handling customer complaints. In the System Availabilitydimension Dana and ShopeePay have the same main topic, namely the digital wallet has not been able to provide quality systems and features so that many users experience failure in top ups and transactions, In the Fulfillment dimension Dana and ShopeePay have not been able to provide satisfaction to customers because there are still many users disappointed with the features provided. And on the Privacy dimension, Dana and ShopeePay have not been able to provide maximum transaction security. From the topic modeling results, it was found that Dana and ShopeePay still did not provide good service.
References
Akerar, R. (2019). Artificial Intelligence for Business. In American Journal of Roentgenology.
Alamsyah, A., & Bernatapi, E. A. (2019). Evolving Customer Experience Management in Internet Service Provider Company using Text Analytics. Proceeding - 2019 International Conference on ICT for Smart Society: Innovation and Transformation
Toward Smart Region, ICISS 2019, October.
https://doi.org/10.1109/ICISS48059.2019.8969828
Alani, M. ., Tawfik, H., & Saeed, M. (2021). Applications of Big Data Analytics. In A. Obinna (Ed.), Big Data Analytics. https://doi.org/10.1201/9781003175711-10
Amalia, A., & Sudiwijaya, E. (2020). Yogyakarta Tourism Promotion using User-Generated- Content Feature. Komunikator, 12(2). https://doi.org/10.18196/jkm.122042
Bahtar, A. Z., & Muda, M. (2016). The Impact of User – Generated Content (UGC) on Product Reviews towards Online Purchasing – A Conceptual Framework. Procedia Economics and Finance, 37(16), 337–342. https://doi.org/10.1016/s2212-5671(16)30134-4
Billyarta, G. W., & Sudarusman, E. (2021). Pengaruh Kualitas Layanan Elektronik (E- Servqual) Terhadap Kepuasan Konsumen Pada Marketplace Shopee Di Sleman Diy.
Optimal, 18(1), 41–62.
DSResearch. (2020). Indonesia Fintech Report 2020 (Issue 47925).
Foster, B. (2020). Pengaruh Servqual Terhadap Kepuasan Yang Berdampak Pada Loyalitas Pengguna E-Wallet (Ovo). BISMA: Jurnal Bisnis Dan Manajemen, 14(1), 48.
https://doi.org/10.19184/bisma.v14i1.10990
Hassanien, A. E., & Darwish, A. (2021). Machine Learning and Big Data Analytics Paradigms: Analysis, Applications and Challenges (A. Darwin (ed.)).
Irwansyah, A., & Mappadeceng, R. (2018). Pengaruh E-Service Quality Terhadap Customer Loyalty Melalui Customer Satisfaction Pada Toko on Line Buka Lapak. J-MAS (Jurnal Manajemen Dan Sains), 3(2), 128. https://doi.org/10.33087/jmas.v3i2.50
Magdalena, A., & Jaolis, F. (2018). Analisis Antara E-Service Quality, E-Satisfaction, dan E- Loyalty Dalam Konteks E-Commerce Blibli. Program Manajemen Pemasaran, Universitas Kristen Petra, 5(2), 1–11.
Malhotra, N. (2018). Consumer Behavior Theories Convergence of Divergent Perspectives with Applications to Marketing and Management. Business Expert Press.
Masrury, R. A., Fannisa, & Alamsyah, A. (2019). Analyzing tourism mobile applications perceived quality using sentiment analysis and topic modeling. 2019 7th International Conference on Information and Communication Technology, ICoICT 2019, 1–6.
Meidita, Y., Suprapto, & Rokhmawati, R. I. (2018). Pengaruh Kualitas Layanan Terhadap Kepuasan, Kepercayaan dan Loyalitas Pelanggan Pada E-Commerce (Studi Kasus : Shopee). Jurnal Pengembangan Teknologi Informasi Dan Ilmu Komputer, 2(11), 5682–
5690.
Mothersbaugh, D. L., & Hawkins, D. I. (2016). Consumer Behavior Bulding Marketing Strategy Thirteenth Edition (13th ed.). McGraw-Hill Education.
https://doi.org/10.2307/1251656
Nurapipah, K. A. F. D. (2021). ANALISIS KUALITAS PELAYANAN MELALUI METODE E-SERVQUAL UNTUK MENGETAHUI KEPUASAN KONSUMEN E-COMMERCE SHOPEE PADA MASYARAKAT KOTA SUKABUMI. JURNAL ILMIAH AKUNTANSI DAN KEUANGAN, 4(2), 133–141. https://doi.org/10.1002/9781119200932.ch9
Ogi, E., Pratiwi1, I., & Yustanti2, W. (2021). Analisis Sentimen Kualitas Layanan Teknologi Pembayaran Elektronik pada Twitter (Studi Kasus Ovo dan Dana). Jeisbi, 02(03), 47–54.
Parasuraman, A., Zeithaml, V. A., & Malhotra, A. (2005). E-S-QUAL a multiple-item scale for assessing electronic service quality. Journal of Service Research, 7(3), 213–233.
https://doi.org/10.1177/1094670504271156
Pinuji, M. F. (2019). Pengaruh User-Generated Content Sebagai Alat Komunikasi Bisnis Terhadap Kredibilitas Informasi Website “Tokopedia.” Inter Script: Journal of Creative Communication, 1(1), 19–33. https://doi.org/10.33376/is.v1i1.347
Rizal, D. A. (2020). Manajemen Pemasran di Era Masyarakat Industri 4.0. In Manajemen Pemasaran.
Rubyanti, D., & Irwansyah, I. (2020). Peran User Generated Content (Ugc) Instagram Pada
Industri Makanan. Mediakom, 4(1), 1–24. https://doi.org/10.32528/mdk.v4i1.3569 Taulli, T. (2019). Artificial intelligence Basics - A Non-Technical Introduction. In
International Carpet Bulletin. Apress.
