TWITTER SOCIAL NETWORK INTERACTION AS CUSTOMER ENGAGEMENT IN COMPETITION FOR E-COMMERCE E-HEALTH PERFORMANCE IN INDONESIA

Background – The presence of e-commerce e-health is a societal solution to health needs in the era pandemic. Social distancing rules cause considerable restrictions on community socialization, so they use social media alternatives to share experiences. That matter causes customers to talk about the performance of digital e-health services and create consumer engagement. Public opinion and ratings on social media become data information analyzed to improve service quality. Aim – This study aims to identify customer conversations on Twitter about e-commerce e-health to see its performance and compare e-health results networks in Indonesia. Design / Methodology / Approach – The research method in this study is the Social Network Analysis (SNA) approach and descriptive meaning to get the results of the formulated goals. Findings – The results of this study showed that pre-pandemic e-health users in Indonesia were more dominant in utilizing psychologist and psychiatric consultations and buying easy and practical medicines. During the pandemic, customers more often used doctor consultations related to symptoms and treatment of Covid-19, drugs and vitamins for self-isolation, supported by online payments. Conclusion - Judging from the proportion of network properties, Halodoc is superior to Alodokter because the number of Halodoc nodes and edges is far more weighted, both from the focus of consulting, medicine, and payments Research implication – This research contributes to e-health companies in Indonesia regarding suggestions for utilizing SNA through customer interaction on social media to improve the performance and competition of e-health services and determine superior strategies. Limitations – This research explores data only from social media Twitter, based on the material for data visualization is only in text. The scope of the study is only in the health business, so it can further develop in other sectors.


INTRODUCTION
Adopting technology in healthcare needs is familiar and has substantially provided good services. Electronic health (e-Health) uses information technology to solve health needs, from prevention to treatment (Crico et al., 2018). During the pandemic, the two most used e-health brands (Surahman et al., 2021). That makes it easier for consumers to get medicines at an appropriate distance, supported by payment services with various methods, such as e-wallets and mobile banking (Suzuki et al., 2020). During the COVID-19 pandemic, people's space for the movement was limited due to large-scale social restrictions (PSBB) to prevent the spread of the virus. The limitations of medical diagnosis are an obstacle due to misinformation about patient symptoms (Maeder et al., 2020). The existence of digital health services such as online doctor consultations, online pharmacies, and noncash payments shows that digital e-health services are a solution to people's needs.
Thus, e-health contributed significantly to many countries through the performance of its digital services during the covid-19 virus outbreak (Alonso et al., 2021).
Service delivery by e-health certainly raises user participation through customerto-customer (C2C) interactions on social media that can support sustainable performance in business (Danarahmanto et al., 2020). Public opinion and assessment of user trust and convenience will create customer engagement (Srinanda et al., 2020) through reactions, interactions, and experiences such as shared uploads and comments (Mukherjee & Banerjee, 2019), (Alamsyah & Utami, 2018). With the number Therefore, careful analysis is needed to investigate social network interactions in performance services to get better results in the future (Carvalho & Medeiros, 2021).
Social network analysis (SNA) explores user interaction relationships, such as uploading, replying, retweeting, and tagging other accounts (Mitei & Ghanem, 2020 Social media platforms with big data include YouTube, Twitter, Instagram, and Facebook (Mai et al., 2020). However, the suitable social media for SNA is twitter, as the licensing process is uncomplicated and access-free, as well as the rapid spread of interactions (Fayjaloun et al., 2021). Tweet data is taken based on user-generated content (UGC), which is content data from interactions originating from users and contains creativity (Fayjaloun et al., 2021), (Saura et al., 2021).  (Kalumata et al., 2021), determinants of e-health (Indriyarti & Wibowo, 2020), and analysis in terms of the quality of e-health applications (Chakraborty et al., 2021), not in terms of user interaction on social media. Likewise, the academic world is more interested in analyzing social networks from the element of financial business (Srinanda et al., 2020) and opinion in education (Lieharyani & Ambarwati, 2022); even so far, there has been no use of SNA in the field of digital health business.
Thus, this study acts further as a filler for the shortcomings of the literature related to ehealth in terms of user interaction networks formed on social media. It also is the novelty of research related to social network analysis methods in the field of health business.
The objectives of this study are twofold.
The first goal is to identify the discussion of

LITERATURE REVIEW Big Data
Big data is extensive information and complex amounts of data that cannot be managed and processed by traditional tools effectively. Big data has a volume at any time with variety and accurate information for further extraction. Thus, big data refers to big social data obtained from social networks (Abkenar et al., 2021).

Social Network Analyst (SNA)
Social Network Analysis (SNA) is part of the Social Computing technique for extracting information on big data, which studies human relations utilizing graph theory. SNA understands social by visually with connected nodes and link lines (edges) on an online social network (Prabowo, 2021).

Customer Engagement
Customer engagement is the result of relational value from the point of view of buyers and sellers, in this case, the extent of the customer's referral value, influence, and knowledge. Customer engagement can turn prospects on existing and valuable social networks, the results of ratings, comments, and reviews, and effective practical (Agnihotri, 2020). and networks (Zhang & Lan, 2022).

Data Collection
Data collection originates from social media After normalization, the dataset saves in CSV form. Finally, filtering with notepad++ removes punctuation marks such as quotation marks and slashes so that they become original words.

Data Analysis And Visualization
Normalization results are processed using word to retrieve important information from the dataset for analysis. The last stage is the utilization of gephi to develop a social network obtained from the phrase bigram to analyze the relationship between its words.
In this study, gephi analyzes the value of the structure of the network of properties formed and visualizes it (Lieharyani & Ambarwati, 2022).

Analysis and Result
This study contains data that compares the results of Halodoc's marketing strategy with Alodokter through interactions on Twitter in consulting, medicine, and payment content.
Researchers made the comparison because there was a drastic change in user interaction before and during the pandemic.      Visualization between node relationships in this network using ForceAtlas2. ForceAtlas2 is a layout algorithm that accelerates graphics to be force-oriented (Lieharyani & Ambarwati, 2022). The researcher chose it because of the use of indirect graphs. The

results of network visualization related to
Halodoc's e-health before the pandemic is in  Part of the medicine's content is users discussing more satisfaction in purchasing drugs, such as "prescription", "delivered", "message", "direct", and "consultation." The emergence of supporting nodes such as "practical", "easy", and "helpful" indicates if the performance of online pharmacy features is acceptable to consumers.
Furthermore, Halodoc customers are talking more about the payment experience, such as "gopay", "discount", "wallet", and "debit" when before the pandemic. The performance of payment services as a support for consultation features and online pharmacies. However, nodes such as "truncated", "failed", and "transaction" also appear if there are problems experienced related to the service.
E-health help customer during the pandemic and even the number of users has increased drastically. Then there is support for payment services to make users share their transaction stories, such as the "gopay", "shopeepay", "discount", "cheap", and "balance" nodes. This node shows that if the more complete and easy payment becomes an attraction for customers, even payment methods with e-wallets get promos and discounts.
The visualized data from figure 5 shows that the discussion of features and benefits received (value) consultation related to the consultation is more dominant related to the pandemic, such as "doctor", "chat", "help", and "fast", which shows that during the pandemic, Alodokter contributed through the performance of digital health services.
The Alodokter drug network in figure 5 discusses the experience of purchasing drugs accompanied by pandemic diagnoses as shown in the nodes "isoman", "cough", "delivery", prescription", and "covid". The nodes indicate that they use alodokter's online pharmacy feature to buy medications with symptoms of covid-19-related illnesses.
Then, network with a discussion of Alodokter payments and is more focused on Visualization with the appearance of "easy", "practical", and "sent" nodes means that users take advantage of the feature to buy drugs delivered with straightforward and practical reasons to the nearest location. A line that provides convenience and speed and determines the closest location to consumers will be a consumer strategy through user experience (Gucen & Hamzah, 2020). As for the visualization during a pandemic, the nodes "pandemic", "vitamins", and "covid" appear.  (Daragmeh et al., 2021).

RESEARCH IMPLICATIONS
This research implies that the e-health startup business industry needs to increase size through active users on Twitter and strengthen customer engagement for feature performance. So that it can be a strategy to develop and grow the e-health business

ACKNOWLEDGEMENT
The author would like to express sincere gratitude to those who have contributed to the completion of this article.