Research on Competitiveness of China’s Social Commerce Enterprises Based on Macro- and Micro-Niche

Imbalance of development and resources allocation can emerge during the rapid expansion of social commerce enterprises, which significantly reduces their competitiveness and sustainability. In this paper, we put forward the concept of social commerce ecosystem and propose a model composed of macro-niche and micro-niche to empirically analyze the performance of different social commerce enterprises in China. The results show that: (1) The social sharing e-commerce subindustry is a monopoly market while the social retail e-commerce sub-industry tends to be a perfectly competitive market; (2) Compared with using only macro-niche or micro-niche, our model shows a more comprehensive insight of enterprise competitiveness, which provides a better niche evaluation for other industries.


Introduction
In the new round of digital transformation, the social characteristics of network information ecology and e-commerce ecological evolution drive the vigorous development of social commerce [1,2]. The social commerce industry is a branch of the e-commerce industry, a new type of e-commerce model that reflects the social value in network socialization. Specifically, it refers to the business activities based on social network service or tools including community, social network and other social media software [3]. In 2019, China's social commerce became the fastest-growing part of e-commerce, with an overall scale of more than 2 trillion yuan (an increase of 63.2%), accounting for more than 20% of online retail and 71% of penetration rate [4]. Besides, many giant enterprises have developed social commerce, such as Alibaba's new app "Taoxiaopu", Jingdong's new independent social commerce platform "Jingxi", TikTok's new social commerce function, and so on. Social commerce has become one of the frontier business models of the digital economy that drives future business and economic prosperity [5,6]. However, many enterprises have rushed for the exit since November 2019. Especially, more than 500 social commerce platforms in Guangzhou and Yiwu closed down in 2020 [7,8]. The main reasons are the fierce competition, fast iteration of the business model and overwhelming flow advantage of giant enterprises. Meanwhile, many social commerce enterprises have problems in the process of rapid expansion, such as unbalanced development and resource allocation [9]. These problems significantly affect the competitiveness and sustainable development of social commerce enterprises, which makes the whole industry a relatively unstable state. Therefore, in order to solve these development problems and improve survivability, it is imperative to evaluate the entire social commerce ecosystem. Wang [35] √ √ Energy vehicle Zhang [37] √ √ E-commerce Zhao and Ling [38] √ √ Wind turbine Qian [40] √ √ √ Port Shi [44] √ √ Home appliance Zhu [45] √ √ √ Logistics Chen [46] √ √ Tourism Yang [47] √ √ √ √ Technology innovation Zhao and Xu [48] √ √ √ International construction contracting Peng [49] √ √ Green technology This paper √ √ √ √ √ √ Social commerce We believe that niche theory is very suitable for the evaluation of social commerce for the following reasons: Firstly, the digital service platform connects many stakeholders and thus is a complex ecosystem [50], and social commerce is a digital platform with service as the core to reflect social value. Secondly, there are sub-industries with significant differences in business models in the social commerce industry and these enterprise groups conform to the concept of population in ecology [51]. In addition, future research trends in the field of value co-creation mainly include the construction of business ecosystem framework and the enterprise niche [52]. As the core research field of value co-creation, social commerce is fully applicable to the prediction. Furthermore, some previous scholars have studied the enterprise niche of e-commerce platform [37]. Therefore, it is appropriate to construct the social commerce ecosystem and evaluate the enterprise niche. At the same time, we put forward the following two research questions:

1.
How is the competitive performance of enterprise roles (macro-niche) and their leading enterprises (micro-niche) in the social commerce ecosystem? 2.
What is the difference between macro-niche and micro-niche of social commerce enterprises? Could the combination of macro-niche and micro-niche evaluation be better than the single niche evaluation?
Through the combination of macro-niche and micro-niche, we expect to discover some interesting "contradictory phenomena", and try to explore the industry essence that has not been paid attention to.
The main contributions of this paper are that we put forward the concept of social commerce ecosystem, propose a model composed of macro-niche and micro-niche to comprehensively evaluate the social commerce ecosystem, and present the comparison between our model and the single niche evaluation. The results show that the social sharing e-commerce sub-industry is a monopoly market while the social retail e-commerce subindustry tends to be a perfectly competitive market. Compared with using the single niche evaluation, our model shows a more comprehensive insight into enterprise competitiveness, which provides a better niche evaluation for other industries.
The rest of this paper is organized as follows: Section 2 provides the research framework of this paper and the construction of the macro and micro index system of social commerce. Section 3 describes data acquisition and processing of social commerce enterprises in China, as well as the niche evaluation methods used in this paper. The results and analysis are shown in Section 4. The conclusions and management implications are presented in Section 5. We respectively take the enterprise roles and the leading enterprises as the research objects, and construct the macro-niche and micro-niche evaluation models to analyze the performance of different sub-industries and leading enterprises. Furthermore, the macroniche and micro-niche are integrated to evaluate the development of social commerce ecosystem comprehensively.

Design of Macro-Niche Evaluation System
According to niche theory and ecosystem view, the social commerce ecosystem is defined as a dynamic balance system with service as the core, reflecting the value transferring of socialization. It contains all participants' interaction in the external environment, including social commerce enterprise groups, businesses, and consumers [3]. The social commerce enterprise groups consist of four main participants: social retail e-commerce (SRE), social sharing e-commerce (SSE), social content e-commerce (SCE), and social e-commerce service provider (SESP) [53]. These participants jointly promote the rapid development of the social commerce industry. According to previous research [38], we consider the macro-niche of social commerce enterprises composed of five niche elements: capital, talent, market, technology and policy, as shown in Figure 2.  We respectively take the enterprise roles and the leading enterprises as the research objects, and construct the macro-niche and micro-niche evaluation models to analyze the performance of different sub-industries and leading enterprises. Furthermore, the macroniche and micro-niche are integrated to evaluate the development of social commerce ecosystem comprehensively.

Design of Macro-Niche Evaluation System
According to niche theory and ecosystem view, the social commerce ecosystem is defined as a dynamic balance system with service as the core, reflecting the value transferring of socialization. It contains all participants' interaction in the external environment, including social commerce enterprise groups, businesses, and consumers [3]. The social commerce enterprise groups consist of four main participants: social retail e-commerce (SRE), social sharing e-commerce (SSE), social content e-commerce (SCE), and social ecommerce service provider (SESP) [53]. These participants jointly promote the rapid development of the social commerce industry. According to previous research [38], we consider the macro-niche of social commerce enterprises composed of five niche elements: capital, talent, market, technology and policy, as shown in Figure 2.

Research Framework
The research framework of the study is shown in Figure 1. We respectively take the enterprise roles and the leading enterprises as the research objects, and construct the macro-niche and micro-niche evaluation models to analyze the performance of different sub-industries and leading enterprises. Furthermore, the macroniche and micro-niche are integrated to evaluate the development of social commerce ecosystem comprehensively.

Design of Macro-Niche Evaluation System
According to niche theory and ecosystem view, the social commerce ecosystem is defined as a dynamic balance system with service as the core, reflecting the value transferring of socialization. It contains all participants' interaction in the external environment, including social commerce enterprise groups, businesses, and consumers [3]. The social commerce enterprise groups consist of four main participants: social retail e-commerce (SRE), social sharing e-commerce (SSE), social content e-commerce (SCE), and social e-commerce service provider (SESP) [53]. These participants jointly promote the rapid development of the social commerce industry. According to previous research [38], we consider the macro-niche of social commerce enterprises composed of five niche elements: capital, talent, market, technology and policy, as shown in Figure 2.  Based on the principles of scientific credibility, concise representation, feasibility and usability, and combined with the previous research [38,39], niche evaluation indexes from five aspects are selected to investigate the niche performance of social commerce enterprises with different roles. The evaluation index of social commerce macro-niche is shown in Table 2. Policy niche represents the strength of social commerce enterprises to obtain policy support. We select the enterprise income tax rate, which is a negative index. The lower the tax rate, the higher the government's support for enterprises.

Design of Micro-Niche Evaluation System
Based on the above evaluation index of social commerce's macro-niche, we consider the social attribute of social commerce users and add the social niche. Social niche represents the level of interaction between social commerce enterprises and users. Due to the high propagation brought by the social network, the social commerce's growth of user scale and transaction scale is more significant, compared with other industries. Therefore, this paper selects user scale, annual growth of users and annual increase in transaction scale to reflect the social niche status of social commerce enterprises. Furthermore, enhancing the richness of the secondary index, the micro-niche evaluation model of social commerce leading enterprises is constructed, as shown in Table 3.

Materials and Methods
With the rapid development of social e-commerce in China, it has begun to take shape. There are well-developed leading enterprises in each sub-industry, which meet the data requirements of macro-niche and micro-niche. Therefore, this paper selects China's social commerce enterprises as examples, expecting to evaluate the entire social commerce ecosystem through our model.

Data Acquisition of Social Commerce Enterprise Roles
In order to evaluate the current situation of social commerce ecosystem, this paper classifies social commerce enterprises by the project descriptions in Tianyancha database. This database is a business query platform under the National Development Fund of China for small and medium-sized enterprises. It contains more than 180 million social entities with more than 90 dimensions of real-time information. Specifically, 1026 social commerce enterprises are selected from this database, filtered by keywords of "social commerce", "social e-commerce" and "community e-commerce" (excluding cancellation and revocation). These enterprises are classified into corresponding roles by the structured content analysis method [54,55]: Firstly, one coder put labels for each enterprise according to its brand description from Tianyancha database, such as social sharing e-commerce corresponds to the following description: group buying, shopping guide, rebate, or coupon, as shown in Table 4. Table 4. An example of a matching process from enterprise introduction to the role.

Pinduoduo
Pinduoduo is a social commerce platform, focusing on the marketing mode of C2B group buying. After the user selects the product, it will send a group invitation to its family or friends through social media. If the number of people meets the requirement, they all can buy the product at the group-buying price. It takes products as the center, guides users to purchase through content and social interaction, and makes the accurate recommendation based on machine algorithms.
Social Sharing E-commerce Inspired by the labels and project descriptions from Tianyancha database, we attempt to describe the enterprise role in social commerce, as shown in Table 5. Table 5. Enterprise roles and descriptions in the social commerce ecosystem.

Enterprise Roles and Descriptions Representative
Social retail e-commerce (SRE) is the primary mode of social value transformation of social commerce, including Business-to-Customer (B2C) mode of traditional e-commerce platform or retailers rapidly transforming and forming scale advantage under their core retail ability, and Supplier-to-Business-to-Customer (S2B2C) mode taking membership e-commerce as a new idea.

B2C: Jingxi, Youyu S2B2C: Yunji, Beidian
Social sharing e-commerce (SSE) relies on the strong relationship to transform social value, mainly in the form of sharing group, supplemented by sharing rebate and obtaining coupons.

Pinduoduo, Taojiji, Peanut diary, Liwushuo
Social content e-commerce (SCE) stimulates consumers' emotional resonance and value recognition through content. Its core competitiveness is continuous, high-quality content output, and it can be divided into community-based SCE focusing on market segmentation and guiding-flow SCE focusing on omni-channel marketing. Then, based on these labels, a second evaluator matches each enterprise to its corresponding role. All authors confirm the final code of each enterprise and discuss the differences among them, which will help to eliminate individual bias.
Finally, we identify four types of general enterprise roles, social retail e-commerce (SRE), social sharing e-commerce (SSE), social content e-commerce (SCE) and social ecommerce service providers (SESP). Information of 1026 enterprises is shown in Table 6. The paid-in capital and number of employees of 1026 social commerce enterprises are from the Chinese National Enterprise Credit Information Publicity System (NECIPS). Their patent counts, software copyright number, work copyright number and website filing information are from the Chinese National Intellectual Property Administration (NIPA). Their financing quota data come from the publicly disclosed news information. The tax rate is formulated according to the Income Tax Law of the People's Republic of China, with 15% tax rate for high-tech enterprises and 20% for small and micro enterprises. We also use web crawler to collect the corresponding App scores and comments of 1026 social commerce enterprises on the four platforms of Huawei App Market, Apple App Store, Xiaomi App Mall and Tencent App Store. In order to ensure the integrity of the data, the final App score of enterprises is obtained by the weighted average of the four platforms' data.

Data Acquisition of Social Commerce Leading Enterprises
We select five leading listed enterprises, namely Pinduoduo, Mogu, Yunji, Youzan and Weimob, which are representative enterprises of SSE, SCE, SRE and SESP, respectively. The data sources of App score, patents and copyrights, and tax rate are consistent with those of macro-niche evaluation indexes. The market cap comes from the stock market's real-time data on 16 July 2020, and the market cap of enterprises listed in Hong Kong's stock market and the American stock market are converted into RMB market cap. The market share is calculated by the proportion of the annual transaction scale of each enterprise. For example, the scale of China's social commerce market would exceed 200 billion yuan in 2019 [56]. Pinduoduo's 2019 annual report shows that its annual transaction scale is 1006.6 billion yuan, so its market share is 50.33%. The data of the user scale, the annual growth of users and the annual increase in transaction scale come from the news information disclosed by enterprises. The indexes of capital niche, the number of employees and the research output-input ratio can be calculated from the 2019 annual report of each enterprise. For example, the research output-input ratio is calculated by the proportion of R&D expenses to operating revenue.

Index Weight
According to the Catastrophe Progression Method, to evaluate the upper indexes, the lower indexes should be first sorted by their importance from large to small [37]. To avoid the subjective ranking, we use the Entropy Method to calculate weights for each index, representing their contribution to upper indexes. The specific calculation method is as follows:

•
Step 1: To ensure the unity of index units, we first normalize the original data: where X ij is the original value of the ith evaluation object under the jth index, Y ij is the normalized value of the ith evaluation object under the jth index.

•
Step 2: Calculate the index weight of the ith evaluation object under the jth index as: where the entropy value of jth index is calculated as: Step 3: Calculate the utility value u i = 1 − e i , the weight of jth index is: according to the principle of entropy additivity, the upper index's weight value in the multi-level structure evaluation index system can be expressed by the sum of the lower index weights.

Niche Value
The normalization formula of the standard catastrophe system model is expressed as where X M is the niche value of upper index M, and X P i represents its corresponding lower niche value sorted by entropy weights. n is the number of lower indexes, which is an integer greater than 0.

Niche Breadth
Social commerce's niche breadth refers to the sum of various resources used by the evaluation object of social commerce enterprises for survival and development. This can be used to measure its diversity of resource utilization, which reflects its potential ability to adapt to the environment and utilize resources. The wider the niche breadth of social commerce, the higher its characterization [47], representing the stronger competitiveness. We use the Shannon-Wiener Exponential Formula [57] proposed by Levin to measure social commerce's niche breadth as: where b i represents the social commerce's niche breadth of evaluation object i, n is the number of evaluation indexes, and P ij = represents the proportion of jth resource used by evaluation object i in its total resources.

Niche Fitness
Social commerce's niche fitness refers to the similarity between the real niche of social commerce and the optimal niche. According to the above social commerce evaluation index system, we construct the measurement model of social commerce's niche fitness as: where F i represents the niche fitness value of the ith evaluation object, and it fluctuates between the interval [0,1]. The larger the value, the more suitable the niche development of the evaluation object is. w j is the weight of each ecological factor, Y ij is the real niche of each evaluation object, h j is the optimal niche of each evaluation object, and α is the model parameter.
• Measurement of the most suitable value h j . With reference to Zhu's assumption, this paper proposes that the social commerce ecological factors follow the normal distribution approximately and the optimal value can be calculated by the sum of the mean and the standard deviation [45].
where µ j is the mean, and σ j is the standard deviation, j = 1, 2, . . . , n. Then h j = µ j + σ j is the optimal value of the ecological factor j. • Measurement of model parameter α. We refer to Li's method of measuring parameter α [58]. Set ε ij = Y ij − h j (i = 1, 2, . . . , m; j = 1, 2, . . . , n), the maximum value, minimum value, and mean value of ε ij are respectively

Niche Overlap
The niche overlap model of social commerce is constructed to analyze the similarity of ecological resources that evaluation objects rely on. The higher the degree of niche overlap, the higher the degree of sharing the same ecological resources, the more intense the competition. In this paper, we use the Pianka Model [59] to measure social commerce's niche overlap and use the normalized data matrix Y ij (i = 1, 2, . . . , m; j = 1, 2, . . . , n) when calculating the niche breadth, then: where O ab is the niche overlap value of evaluation object a and evaluation object b, and the value range is [0,1]. P aj represents the proportion of the jth resource used by evaluation object a in its total resource utilization, P bj represents the proportion of the jth resource used by evaluation object b in its total resource utilization. We sum the secondary index of enterprises in the same role in social commerce to obtain the original data, where each role represents a population. The weights of primary and secondary indexes evaluated by the Entropy Methods are shown in Table 7. According to the Catastrophe Progression Method, the four enterprise roles' niche values are calculated, as shown in Table 8. It is found that social retail e-commerce (SRE) occupies the first place in the comparison of the four enterprise roles. Three of the five primary niche indexes occupy the first place, but the market niche performance is poor. In the secondary niche index comparison, its financing quota is lower than that of social sharing e-commerce (SSE) and social ecommerce service providers (SESP). Meanwhile, its App score is lower than that of social sharing e-commerce (SSE) and social content e-commerce (SCE). It can be seen that SRE, as the primary form of social commerce-"socialization of retail e-commerce", has a low entry threshold and the largest number of enterprises. In the early stage of the development of the social commerce industry, it has occupied a dominant position in all aspects. However, the development of the social commerce industry has entered a new stage. The investment funds are inclined to the SSE, a relatively new social commerce model. The industry has begun to realize the importance of SESP's innovation and development. Simultaneously, the consumer agglomeration phenomenon of SRE is significant, and consumers have a low evaluation of the uneven application.

Results and Discussion
In the comparison of the primary niche index of SSE, the market niche ranks first. Its capital niche and technology niche also have good performance, second only to SRE, while the policy niche performs poorly and ranks the last. When comparing the secondary niche index, the niche performance of financing quota and App score is the best. Except for the poor performance of software copyright and tax rate, the other indexes perform well on the whole, which is consistent with the second ranking in the total niche comparison. SSE led by "Pinduoduo" has developed rapidly in recent years. Although it cannot compete with SRE in the number of enterprises, its leading effect is significant, which drives the rapid development of the SSE sub-industry.
Although SCE is not dominant in total niche comparison, its policy niche and market niche perform well, especially the niche of the tax rate, which is the best among all roles, showing the greatest policy support. The values of the capital niche and talent niche are the lowest, and there is a large space for improvement.
SESP ranks second or third in comparing the five primary indexes except for market niche, which is consistent with its total niche performance. The secondary index does not perform well in terms of App score and patent counts, ranking the last. The reason is that SESP, as a supporting role in the social commerce industry, mainly focuses on enterprise services, and its technical basis is mostly from the e-commerce and communication industry. The development space of patent innovation combined with the characteristics of social commerce is vast.

Niche Breadth
We calculate the proportion of the jth resource used by each enterprise role in its total resource utilization P j . The niche breadth of four enterprise roles in social commerce can be calculated by substituting it into the Formula (6), as shown in Table 9. Niche breadth values of the four enterprise roles in social commerce are more significant than 1, indicating that each role has a good performance in resource utilization. The niche breadth value of SRE is greater than 2, ranking the first, while SSE and SESP are ranked second and third with a slight gap. SCE is at the bottom of the list. The ranking and numerical differences of niche breadth of the above social commerce enterprise roles are consistent with their niche values.

Niche Fitness
The normalized evaluation index data Y ij is substituted into Formula (7) to calculate model parameter α and optimum value h j . Combined with the index weight w j which is calculated previously, the social commerce's niche fitness is measured, as shown in Table 9. The four enterprise roles are sorted. In this paper, we refer to Zhu [45] to set five levels of fitness: when 0 ≤ F i ≤ 0.4, unfitness; when 0.4 < F i ≤ 0.5, low fitness; when 0.5 < F i ≤ 0.6, fitness; when 0.6 < F i ≤ 0.8, moderate fitness; when 0.8 < F i ≤ 1, high fitness. This paper finds that social commerce's macro-niche fitness is relatively high, and there are no unfitness and low fitness. Except for the niche fitness of SSE, the other three enterprise roles' niche fitness is moderate. The lower fitness of SSE may be due to the unbalanced resource allocation caused by the rapid development of the shopping e-commerce platform led by Pinduoduo in a short period. In general, social commerce's macro-niche meets the needs of social commerce development. Its resource allocation is relatively reasonable, but there is still room for progress.

Niche Overlap
The proportion P j of the jth resource used by each enterprise role in its total resource utilization is substituted into Formula (8) to calculate the niche overlap, as shown in Table 9. It is found that the changing range of the four enterprise roles' niche overlap in social commerce is [0.2990, 0.6629], and the proportion of low, medium and high competition is 1:3:2. In general, there is moderate competition in the niche overlap of enterprise roles. That is, the similarity degree of enterprise roles in resource utilization is moderate, and the degree of competition is moderate. The reason is that there is a big difference in the business model among four enterprise roles. The reason for the high niche overlap between SESP and SCE providers is that some service providers undertake the function of guiding to the platform. To a certain extent, they have some business overlapping with the guidingflow SCE. The niche overlap of SRE and SESP is moderate, which is since most of the two enterprise roles are transformed from traditional retail platforms or e-commerce service providers and their essential resources are similar.

Macro-Niche Analysis
Based on the niche breadth and niche fitness of each enterprise role, we find that social retail e-commerce (SRE) and social sharing e-commerce (SSE) have strong survivability and weak resource allocation ability, while social e-commerce service providers (SESP) and social content e-commerce (SCE) have weak survivability and strong resource allocation ability. We refer to this as the inversion phenomenon of survivability and resource allocation ability in the social commerce industry.

Niche Value
In this paper, the Entropy method is used to calculate the weight of each evaluation index of micro-niche, as shown in Table 10. According to the Catastrophe Progression Method above, the niche values of evaluation indexes of the five leading enterprises are calculated, as shown in Table 11. Among them, Pinduoduo occupies the first place in the total niche comparison of five social commerce enterprises, which is far higher than the other four. In the comparison of the primary indexes, the market niche, social niche, talent niche and policy niche of Pinduoduo rank first, while the capital niche and technical niche are not dominant. Its capital niche lags behind Mogu and Weimob, and its technical niche is lower than that of Mogu. From the view of secondary evaluation indexes, the App score and asset-liability ratio rank the last, and the profit growth ranks the fourth. The reason is that the low-quality and low-price product gathering in Pinduoduo platform inevitably leads to the problem of word-of-mouth differentiation, which makes its App score low. Its product gathering characteristic limits the space of Pinduoduo's profit growth.
Compared with the five social commerce enterprises, Mogu ranks fourth in the total niche. Among all its primary niches, the market niche, social niche and talent niche rank the last. In contrast, the policy niche, capital niche and technology niche rank first. From all indexes of Mogu, we find that its niche value is either the largest or the smallest. It shows the polarization problem and the uneven development of Mogu in the social commerce ecosystem. In the future, without damaging the dominant niches, efforts should be made to make up for the inferior niche, such as absorbing outstanding talents and strengthening the social interaction with users.
The total niche of Yunji ranks the last, and its social niche ranks the second while its other primary indexes are basically the last but one. The annual growth of users is the largest, while the other secondary indexes are almost the second last.
Youzan's performance of policy niche and capital niche is deficient in the primary index comparison, ranking the last. It is worth mentioning that Youzan has the most extensive research output-input ratio. It can be seen that despite the lack of funds in Youzan, it attaches great importance to R&D.
Weimob ranks first or second in most primary indexes, which is consistent with its total niche performance of the second ranking. However, the performance of the social niche is poor, ranking the fourth, and the secondary index of user scale is even the last. Besides, among the secondary indexes, the research output-input ratio is the lowest. It can be seen that Weimob has advantages in most resources, but it still needs to strengthen user interaction and increase R&D investment.

Niche Breadth
According to the above niche breadth model of social commerce, the niche breadth of five leading enterprises is calculated, as shown in Table 12. From Table 12, the niche breadth of the five leading enterprises of social commerce is about 2, which indicates that they all have nice performance in resource utilization. Specifically, the niche breadth of Pinduoduo is 2.3976, ranking first. It is followed by Weimob, 2.2 and Mogu, 2.081; Youzan and Yunji are lower than the first three, which are 1.9298 and 1.8368 respectively. The ranking and numerical differences of niche breadth of the above five leading enterprises in social commerce are consistent with their total niche values.

Niche Fitness
The niche fitness of five listed leading enterprises is measured according to Formula (7) and are ranked, as shown in Table 12. Under the five levels of fitness, the niche fitness of Yunji is the highest, reaching moderate fitness; the niche fitness of Mogu is the lowest, which is low fitness; the niche fitness of the other three leading enterprises all belong to fitness. Combined with the above analysis of the niche breadth of leading enterprises, we find that the niche breadth of Yunji is the lowest, while its niche fitness is the highest, which indicates its "small and fit" business model. The niche fitness of Mogu belongs to low fitness, which is related to the severe polarization of various resources, and the low fitness of Mogu is consistent with its low niche breadth.

Niche Overlap
According to the above niche overlap model of social commerce, we can calculate the niche overlap of five listed leading enterprises, as shown in Table 12. The changing range of five leading enterprises' niche overlap is [0.3169, 0.6881], and the proportion of low, medium and high competition is 0:7:3. In general, the competition among these five leading enterprises is moderate. The reason is that the five leading enterprises of social commerce belong to four different enterprise roles. Moreover, there is no evident homogenization among different enterprise roles' business model. The reason for the low niche overlap of Youzan and Weimob, which both belong to SESP, is that Weimob mainly relies on WeChat system to provide omni-channel mini programs as e-commerce solutions for businesses. Therefore, Weimob mainly adopts the agent mode for promotion. However, Youzan mainly adopts the free trial and the local promotion mode to attract businesses. These are apparent differences between the two business models.

Micro-Niche Analysis
Based on the niche breadth and fitness of leading enterprises, we can summarize their survivability and resource allocation abilities as follows: Among them, Yunji (SRE) has the weak survivability and the strong resource allocation ability. Pinduoduo (SSE), on the contrary, has strong survivability and the weaker resource allocation ability. The resources relied on by Mogu (SCE) are seriously polarized. Youzan and Weimob (SESP) adopt differentiation strategies, carving up market share "peacefully".

Macro and Micro Combination Analysis
This paper makes a comparative study on the macro-niche according to the enterprise roles and the micro-niche represented by leading enterprises. The micro-niche of SESP is computed by averaging the niche values of Youzan and Weimob. The integrated results are shown in Table 13. We can consider the macro-niche value of the sub-industry as its average level to a certain extent and hence show both macro-niche and micro-niche in the same graph as Figure 3. We can consider the macro-niche value of the sub-industry as its average level to a certain extent and hence show both macro-niche and micro-niche in the same graph as Figure 3. In Figure 3, the bars with different colors represent the micro-niche index values of corresponding sub-industries, and the dotted lines represent the corresponding macroniche values. From Table 13 and Figure 3, the niche breadth of SRE is the first, but its niche fitness is low. Surprisingly, Yunji (SRE) ranks the last in niche breadth and ranks first in In Figure 3, the bars with different colors represent the micro-niche index values of corresponding sub-industries, and the dotted lines represent the corresponding macroniche values. From Table 13 and Figure 3, the niche breadth of SRE is the first, but its niche fitness is low. Surprisingly, Yunji (SRE) ranks the last in niche breadth and ranks first in niche fitness. It is related to the large number of SRE enterprises, indicating fierce competition. Yunji's rationality of resource allocation is high, but it does not occupy an oligopoly position in the SRE sub-industry. Therefore, we believe that the SRE sub-industry tends to be a perfectly competitive market. Besides, the niche breadth of Pinduoduo (SSE) ranks first, and its competitiveness is far higher than the average industry level and other leading enterprises. However, the niche breadth of SSE still ranks second. It can explain, to a certain extent, that Pinduoduo is dominant in the SSE sub-industry, and the development level of other SSE enterprises is trivial. Therefore, we suggest that SSE sub-industry is a monopoly market. Moreover, the niche fitnesses of Mogu, Youzan and Weimob are lower than those of their average industry levels, so their rationalities of resource allocation need to be strengthened.
The niche overlap of social commerce's macro-niche is similar to that of micro-niche, which can be seen in that their niche overlap ranges are similar, [0.2990, 0.6629] and [0.3169, 0.6881] respectively. The competitions between most of the evaluation objects are moderate, no matter from macro-niche or micro-niche evaluation.
In this paper, macro-niche and micro-niche with niche value, breadth and fitness are integrated as the comprehensive competitiveness evaluation of each sub-industry in social commerce, as shown in Figure 4.  From Figure 4, it is found that the order of the hexagon area is SRE, SSE, SEEP a SCE from large to small. Among them, the last three are mutually included in turn, wh SRE and the other three have significant differences in the performance of each dimensi On the whole, we suggest that the competitiveness of SRE and SSE is in the leading po tion, while the competitiveness of SESP is stronger than that of SCE.
Due to the significant individual differences among the leading enterprises, it is n representative to only use micro-niche to evaluate the competitiveness of the social co merce industry. Therefore, we only compare the results of our model and the sin macro-niche evaluation. The radar chart of macro-niche evaluation is shown in Figure   Figure 4. Comprehensive niche performance of social commerce.
From Figure 4, it is found that the order of the hexagon area is SRE, SSE, SEEP and SCE from large to small. Among them, the last three are mutually included in turn, while SRE and the other three have significant differences in the performance of each dimension. On the whole, we suggest that the competitiveness of SRE and SSE is in the leading position, while the competitiveness of SESP is stronger than that of SCE.
Due to the significant individual differences among the leading enterprises, it is not representative to only use micro-niche to evaluate the competitiveness of the social commerce industry. Therefore, we only compare the results of our model and the single macro-niche evaluation. The radar chart of macro-niche evaluation is shown in Figure 5.
On the whole, we suggest that the competitiveness of SRE and SSE is in the leading position, while the competitiveness of SESP is stronger than that of SCE.
Due to the significant individual differences among the leading enterprises, it is not representative to only use micro-niche to evaluate the competitiveness of the social commerce industry. Therefore, we only compare the results of our model and the single macro-niche evaluation. The radar chart of macro-niche evaluation is shown in Figure 5. From Figure 5, we can suggest that SRE is in the leading position, SESP in the second place is slightly superior to SSE, and SCE has the worst performance among them. However, our model further considers the significant influence of leading enterprises. As a result, SSE surpasses SESP in the second place due to the excellent performance of Pinduoduo (SSE). Therefore, we believe that our model can correct the one-sided evaluation From Figure 5, we can suggest that SRE is in the leading position, SESP in the second place is slightly superior to SSE, and SCE has the worst performance among them. However, our model further considers the significant influence of leading enterprises. As a result, SSE surpasses SESP in the second place due to the excellent performance of Pinduoduo (SSE). Therefore, we believe that our model can correct the one-sided evaluation from the single niche perspective and provide more information about the competitiveness of the social commerce industry.

Conclusions
In this paper, we put forward the concept of social commerce ecosystem and propose a model composed of macro-niche and micro-niche to empirically analyze the performance of different social commerce enterprises in China. Conclusions can be drawn as follows: (1) The evaluation of social commerce macro-niche shows that the survivability of social retail e-commerce (SRE) and social sharing e-commerce (SSE) is leading, while the resource allocation rationality of social e-commerce service providers (SESP) and social content e-commerce (SCE) is better, which shows an inversion phenomenon. (2) The evaluation of social commerce micro-niche shows that the Pinduoduo (SSE) has superior survivability, Yunji (SRE) and Mogu (SCE) have balanced and polarized resource allocation respectively, and Youzan and Weimob (SESP) adopt differentiation strategies. (3) The combination of macro-niche and micro-niche shows that SSE is a monopoly market, while SRE tends to be a perfectly competitive market. This integrated method shows more comprehensive results, which is better than the single niche evaluation method.
Based on these conclusions, aiming at the individual problems of leading enterprises and the common problems of each enterprise role in social commerce, this paper provides the following suggestions: (1) A single supervision mode is not conducive to the sustainable development of the social commerce industry. Targeted and effective industrial policies should be adopted for different characteristics of social commerce sub-industries. For example, the SSE sub-industry belongs to a monopoly market, and hence, incentive policies should be introduced to vigorously support the development of small and medium-sized enterprises and encourage benign competition of the industry. On the other hand, SRE tends to be a perfectly competitive market and subdivision supervision policies should be adopted to prevent the risk of vicious competition. (2) In the rapid development of social commerce enterprises, the imbalance of resource allocation is severe. When the enterprise develops to a certain stage, the development focus should be shifted from "speed" to "balance". For example, the polarized resource allocation of Mogu (SCE) seriously restricts its development sustainability. Therefore, in the future, its attention should be paid to absorbing industry talents and enhancing the interaction between products and users. (3) Social commerce enterprises should pay special attention to the development of differentiation strategy, so as to reduce the overlap between the enterprise roles and even between the leading enterprises in the same role. For example, Youzan and Weimob, both in the SESP sub-industry, adopt different business models and profit models. Therefore, when seizing the market share, they could avoid the fierce conflict, which is beneficial to the sustainable development of the whole sub-industry.
Although this study has some innovation among studies of the same type, it still has some limitations. For example, due to the short development time of social commerce, research can only be carried out from the static perspective, without considering the change of indexes over time. The second limitation is the lack of data of small and medium-sized enterprises in social commerce, which makes it hard to cover a large number of macro-niche evaluation indexes, especially indexes of social attributes. Besides, this study is based on the China region, which is representative of the development of social commerce, but it can still cause a lack of model universality. The future research will further consider the evolution law of social commerce niche from a dynamic perspective and extend the model to other countries or regions.
Author Contributions: Conceptualization, Y.W. and X.C.; methodology, X.C.; validation, X.C., Y.W. and R.Z.; formal analysis, X.C.; data curation, R.Z.; writing-original draft preparation, X.C.; writingreview and editing, Y.W.; visualization, R.Z. All authors have read and agreed to the published version of the manuscript. Data Availability Statement: Data available on request due to restrictions eg privacy or ethical. The data presented in this study are available on request from the corresponding author. The data are not publicly available as the data also forms part of an ongoing study.