Internationalization Pace, Social Network Effect, and Performance among China’s Platform-Based Companies

Abstract

Platform-based companies are rapidly emerging and occupy major positions in the global market. In addition, the business model that relies on digital technology makes platform-based companies sustainable. This paper empirically examines the impact of internationalization breadth on corporate performance and explores the moderating role of the platform enterprise network effect and innovation investment. Annual panel data from listed Chinese platform-based companies between 2012 to 2020 were used to also verify the scale of overseas companies’ association and social networks and the moderating effect of centrality and innovation research and development (R&D) investment on these two factors. The results indicate that the irregular internationalization rhythm of platform-based companies was negatively correlated with corporate performance, and the association network and the social network centrality of overseas companies had a positive moderating effect. Innovative R&D investment could help platform-based companies expand overseas to improve their corporate performance, and its moderating effect was more significant for platform companies in the early stages of internationalization. In terms of the internationalization pace, overseas network construction, and innovative R&D investment strategies, this paper provides valuable suggestions for platform-based companies to expand internationally and improve their corporate performance.

1. Introduction

Platform-based companies are those that have extremely high resource organization and mobilization abilities in a relatively open and orderly manner to achieve effective resource allocation. They are of great help to China in conducting corresponding anti-pandemic work and resuming work and production. According to the global listed company stock market capitalization in the New York Stock Exchange from the Capital IQ database at the end of 2021, 6 of the top 30 global companies were platform-based companies, including Alibaba [1]. In recent years, Chinese digital information services, cross-border e-commerce, and other platform-based companies have continued to develop in the global market. These companies have expanded overseas and built a social network system that is conducive to internationalization by integrating relevant domestic and foreign information resources. The next step for Chinese platform-based companies is to set up overseas branches and conduct global expansion. Platform-based enterprises often have capabilities that other traditional enterprises lack, such as non-competitiveness, network effects, and scale effects. These characteristics provide platform enterprises with the ability to radiate globally. In addition, platform-based enterprises can be used to obtain resources more effectively to increase their ability to become sustainable. While traditional enterprises are often subject to geographical and national policy restrictions during the process of overseas expansion, platform-centered enterprises benefit from globalization, can provide large-scale, high-quality, customized, and innovative service capabilities, and thus can develop into a giant economic power. Due to their own network effect and fast development, platform-based companies are often characterized by “aggressiveness” and “leapfrogging” when entering international markets, and it has become a trend for them to rapidly occupy these markets, in which they consolidate their leading position with emerging technologies and innovative development. However, China’s platform-oriented enterprises have faced various degrees of difficulties and bottlenecks in overseas business expansion. In particular, due to the rapid and aggressive placement in India, South Africa, and other regions that offer little margins, an expansion plan eventually had to be terminated. In recent years, for example, Byte Spring first explored the Indian market, with good results at the early stage, but the subsequent excessive investment for an accelerating expansion resulted in an overall loss of $12.2 billion in 2018.

Whether the irregular pace of the internationalization of platform-based firms can help these firms achieve their higher performance aims needs to be further explored. In the booming digital economy, new forms of the international expansion of platform-based firms are emerging, which have not yet been theoretically explained. Vermeulen and Barkema [2] divided the internationalization process into three dimensions, namely, speed, pace, and scope. It is generally accepted that international expansion is complex and that different processes and modes of expansion (e.g., scope, speed, and pace) can lead to significant differences between firm performance and cost efficiency. The extant literature consistently shows that research on the internationalization process and firm performance is relatively mature. However, most scholars have focused on the two dimensions of internationalization speed and scope, and there is a lack of research on the dimension of internationalization pace, which is not supported by empirical data. Traditional firms are the research objects of the existing theories on internationalization and performance, but the current rise of platform-based firms has created an urgent need for internationalization theories and an understanding of the impact mechanisms of internationalization and performance that are applicable to the characteristics of platform firms.

This paper investigates the internationalization behavior of platform firms based on a dynamic perspective and examines the mechanisms that influence the internationalization rhythm and performance decisions of platform firms in the following three aspects. The objective of this study is to find specific mechanisms for platform-oriented firms in their internationalization performance and to provide theoretical and empirical basic data for their internationalization strategies and performance decisions. First, we depart from the traditional research methodology on internationalization and examine the impact of the internationalization process on the performance of platform-based firms from the perspective of the internationalization of these firms by focusing on the internationalization rhythm dimension. Second, we combine the network effect with the social network and platform network effects to investigate their moderating effects on the relationship between internationalization rhythm and performance and explore the effects of network building in the internationalization process of platform-oriented enterprises. Third, from the perspective of corporate innovation, we investigate the moderating effect of innovation investment on the relationship between internationalization rhythm and performance and explore the role of innovation R&D investment in the internationalization process of platform firms.

2. Theoretical Analysis and Research Hypothesis

2.1. Platform Enterprise and Platform Ecology

Domestic and international research on platform-based enterprises began with the concept of a “platform organization”, which was first introduced by Ciborra [3]. In addition, Danciu [4] pointed out that platform-based organizations have the ability to better adapt to their constantly changing sustainability challenges. A platform-based organization can create a flexible mix of resources, practices, and structures in the context of emerging business opportunities and challenges. Xu and Zhang [5] regarded a platform as a virtual or real space that can facilitate transactions between two or more customers and highlighted that platform-based enterprises based on the value system can combine corporate value and customer value. Through an in-depth study of bilateral market theory, Rochet and Tirole [6] concluded that platform firms do not provide goods or services directly to buyers and sellers; instead, the “platform” makes buyers and sellers mutually attractive, and parties can trade on the platform at reasonable rates. Based on the theoretical foundation of platform enterprises laid by Rochett and Tiroleo [6], some scholars have summarized the characteristics of platform enterprises. First, platform enterprises have the characteristics of a “bilateral structure” or “multilateral structure”, which connects two or more buyers or resource providers. Second, platform enterprises have network externalities or network effects. One party in a platform-based network is influenced by other parties and thus benefits from the characteristics and scale of the other parties, as reported by Boudreau and Jeppesen [7]. Third, platform enterprises have the feature of openness. Platform firms openly interact with other market players to create a system that affects their own opportunity identification, i.e., network effects or network externalities. In addition, network effects or network externalities can be classified into direct and indirect types. Direct network externality refers to the value effects associated with the number of users of products and services, i.e., same-side impact. Indirect network externality refers to the influence on one platform participant by the market size of other participants, i.e., the degree of participation in the platform depends on the size of other participants in the platform. Chen [8] argued that network externalities can impact the effective expansion and scale development of platforms, thus laying a theoretical foundation for studies related to platform market organization. Network effects are also assessed into the following five categories: acquirer, competitor, participant, market effect, and economic effect indicators.

Platform-based firms do not exist in isolation but are embedded in a platform ecosystem. Iansiti and Levien [9] elaborated on the broad ecosystem, which they suggested is a group of individuals and organizations interacting with each other, in which the participants develop collaboratively, play their respective roles, develop their capabilities, and align with the goal direction of the central organization to which they belong. Based on the platform context, Adner and Kapoor [10] defined a platform ecosystem as a network with complementarities in which the platform itself and its complementary nature can work together to enhance the value of the platform. There are both internal and external platform-based ecosystems [11]. While an internal platform-based ecosystem is defined as a set of resources in a common structure from which a company can efficiently develop and produce a range of derivative products, an external platform ecosystem is an innovative business ecological unity formed by the products, services, or technologies of an external innovator base that together can develop their own complementary products, technologies, or services. Li et al. [12] compared the governance of platform-centric ecosystems and network multinationals. The theory of ecosystem-specific advantage (ESA) was proposed to extend the outreach of ecosystem theory and consider an ESA as a type of firm-specific advantage (FSA).

2.2. Impact of the Internationalization Pace of Platform Firms on Firm Performance

The international expansion process of firms is complex, and the differential internationalization process can lead to significant differences in the performance of firms in this process. Vermeulen and Barkema [2] suggested that internationalization rhythm describes the varied regularity and consistency of firms’ international expansion, which results in different internationalization rhythms due to the differences in these patterns.

In selecting indicators for the internationalization pace variable of a firm, Zhong et al. [13] used the change in a firm’s overseas revenue as a measurement variable. The inverse natural logarithm of the standard deviation of the regression coefficient of the change in revenue over a certain period of time was chosen for their empirical study, in which the larger the value of this indicator, the more irregular the internationalization rhythm of the firm. In terms of the relationship between internationalization rhythm and firm performance, Li [14] considered rhythm as an indicator of the speed of foreign investment and argued that too fast a pace of foreign direct investment is detrimental to industrial upgrading. Zeng et al. [15] argued that with an irregular international expansion rhythm, the future development of a firm cannot be effectively predicted, and indicators such as business risks are difficult to assess in advance. Based on absorption theory, a high-risk, irregular international expansion model would bring more difficulties and challenges to a firm, leading to a lower level of performance. A regular pace of internationalization would enable the firm to increase its absorption and learning capacity. Through organizational learning, regularly paced firms can transform the problems and failures encountered during their international expansion into their own knowledge and capabilities, thus improving their development. Hashai et al. [16] proposed that firms can increase their own predictability during their regular internationalization, and thus, firms can predict risks in international expansion through past organizational experience and operations and improve efficiency and their own corporate management capabilities to reduce management costs.

2.3. Impact of Social Network Effects on Firm Performance

Platform-based firms are connected to each other in a complex network structure. In previous studies on clustered enterprise networks, the size, density, relationships, and location of a network are usually used to measure its structure. Network density, which reflects the degree of connectivity among the nodes in a network, has a significant impact on the firms that are connected to the network. This index is generally measured by the ratio of the actual number of relationships with network associates and the maximum number of possible relationships. The higher the number of actual relationships with other firms, the higher the network density of the cluster in which the firm is located. Gupta et al. [17] showed that firms with high-density networks and high information sharing can facilitate easy exchanges of information and resources among participants, and therefore promote the division of labor and cooperation between them. At the same time, a high network density has a disincentive effect on cooperation and resource sharing among network participants for those who are “free riders” and opportunistic. Network size is generally defined as the total number of networks within a company in an internet relationship cluster and is measured by the number of partners with whom the company has a direct partnership. In their study of the network size of companies, Baum [18] highlighted that a large network size increased the amount of information available to the company and enriched the diversity of all network relationships within the company through the expansion of the total amount of information, thus allowing access to a large amount of heterogeneous information. According to the structural hole theory proposed by Burt [19], a firm’s position in a network is more important than the strength of its relationships and determines its information, resources, and power. Network position is generally used to evaluate the importance of a company in a relevant network and is also an important and significant social asset. When a firm is in a relatively important economic network position, it may be more helpful for the firm to have access to a large number of network resources, enhancing the firm’s control and decision-making advantage. There is a paucity of existing research on the relationship between network effects and firms’ internationalization performance. In terms of theoretical development and empirical evidence, the relationship between social networks and firms’ internationalization performance has been examined by scholars. Through empirical analysis, Zhou et al. [20] found that social networks act as mediating variables that link the internationalization process with firm performance. Li et al. [12] reported that firms can use overseas social platforms to locate the negative effects of a lack of market information and thus better meet the individual needs of overseas users, thus reducing the negative effects of irregular rhythms on firms’ internationalization. The internationalization of platform-based enterprises and their ecosystems is beyond the scope of theories about the internationalization logic of traditional multinational companies; hence, new theoretical thinking is needed to understand the digital platform-centered ecosystem and its internationalization. From the perspective of the various agents participating in a platform ecology, Eisenmann et al. [21] proposed that effectively distinguishing between platform providers and platform owners can be helpful for analyzing the competitive strategies of platforms more accurately. A platform ecology includes four types of subjects: demand-side users, supply-side users, platform providers, and platform owners. Complementary goods and platform users, together with the platform firms themselves, constitute a relational network. In addition, an effective strategy helps the company stand out from the competing entities in the market [22]. Therefore, firms in a platform ecology develop competitive strategies and create mechanisms to attract a large number of complementarians, enrich their diversity, expand their network size, and establish mutual ties through transactional behaviors to make the platform ecology network more stable. When the network effect is strong, the stable position of the network ecosystem built by the platform increases the cost of belonging to the platform; thus, the dependence of platform users becomes stronger, so platform companies occupy a key position in the market with their strong ecological network [23]. Based on existing research on network effects and firm performance with an ecosystem perspective, Nambisan et al. [24] suggested that these studies usually take the platform ecosystem as the subject of study; however, it is easy to consider the ecosystem as an external environment in which the platform firm is located, and an analysis of how an individual firm can improve performance and enhance initiative in the ecosystem is lacking.

2.4. The Impact of Innovation Investment on Firm Performance

Investment in R&D is an important way for enterprises to enhance their core competitiveness, which determines their effectiveness and development prospects. Most scholars believe that R&D investment is positively related to innovation performance. Garner et al. [25] empirically concluded that R&D investment has a significant positive effect on the speed of innovation and thus improves the innovation performance of enterprises. However, there is a trade-off between the value creation effect and the value distribution effect generated by a competitive strategy. In terms of the value creation effect, some scholars suggest that a more open platform competition strategy results in lower and more open technology standards and technological barriers so that complementary players can have easier access to such a platform ecological network, thus attracting more complementary players to the platform and increasing the technological and economic effects of it. In contrast, a relatively closed platform competition strategy strengthens the protection of intellectual property rights and the control of the organizational structure, which eliminates lower quality complementarities when they enter the platform ecology to screen out the high-quality complementarities and improve the complementarity and coupling of the platform and the complementary platforms, thus enhancing the technical performance of the platform ecology. Domestic and international studies show that the research on the internationalization process and firm performance is more mature, but most scholars only focus on the two dimensions of internationalization speed and scope, and there is a lack of research on the dimension of internationalization rhythm. At the same time, the existing theories on internationalization and performance are based on traditional enterprises, but the rise of platform enterprises in recent years has created a pressing need for internationalization theories that are applicable to the characteristics of platform enterprises and the impact mechanisms of internationalization and performance. In this paper, we chose internationalization pace as the dimension through which to explore the relationship between the internationalization pace and the performance of platform firms. From the perspective of network effects, we explore how the platform network effects and social network effects of firms moderate the relationship between internationalization rhythm and the performance of platform firms.

3. Research Hypothesis

3.1. Relationship between Internationalization Pace and Performance of Platform-Based Companies

The born global theory refers to companies with the potential to accelerate internationalization and a global market perspective [26]. Therefore, born global firms have strong competitive advantages over traditional large-brand companies with a larger scale and more stable businesses, reflecting business profitability and a rapid business growth model that differs from those of traditional companies. As emphasized in the modified Uppsala model, the network position is an indispensable variable for studying the modern internationalization process, in which the firm is in a multiparticipant business network that includes participants in various interdependent relationships. Internationalization is seen as a result of a firm’s actions to strengthen its network position. Since the network is borderless, the distinction between the firm’s entry and expansion in foreign markets is weakened. Therefore, strengthening the firm’s position in the network for internationalization is more important than the traditional effort of market entry (i.e., overcoming barriers). Platform-based companies are mostly dependent on the digital infrastructure of the internet to be able to communicate and collaborate and to transact and create value through online digital business models. They can internationalize their business activities at a faster rate of expansion from the very beginning of their existence. The online automation, network effects, and flexibility and autonomy of the network are more advantageous than those of traditional companies. Metcalfe’s Law shows that growth in the number of users leads to growth of the network value. Due to the existence of network externalities, as the number of users increases, the expansion rate of the network also increases; when the total number of users exceeds the critical point, a clear dominant position is formed. When platform enterprises enter the international market, they make use of their own network advantages to quickly achieve global large-scale expansion, the choice of target markets and entry methods gradually loses importance, and the influence of local partners on the globalization of platform-based enterprises gradually increases. Compared with the traditional independent variables of platform firms, platform firms have a lower outsider disadvantage in internationalization because they have the following variables: network effect of corporate performance, innovation input, and the social network effect. Therefore, in their internationalization mode, platform-based firms are more aggressive and exert a wider leapfrogging effort to achieve the scale effect more quickly through network expansion. The internationalization of irregular platform firms, which initially stems from the natural network advantage of platform firms, may lead to a period of rapid foreign investment and a period of slowed foreign investment due to conservatism, resulting in an unstable pace of foreign expansion. In the process of platform enterprises entering international markets, from the perspective of absorption capacity, enterprises with a stronger absorption capacity can more easily acquire external knowledge, and thus promote innovation. When multinational enterprises conduct overseas business, they need to absorb and digest newly acquired knowledge and experiences. From the perspective of “time compression diseconomies”, platform-based enterprises must carefully choose the scale and pace at which they enter the international market. Although enterprises can gain many benefits from internationalization, this does not mean that the faster the pace of enterprise internationalization, the better the outcome for the enterprise. For platform-based enterprises, new systems, structures, and networks in the internationalization process often need a long time to break in before they work well. The online circulation and transactions of digital platforms involve uncertainty and higher costs at the level of trust and understanding than those of offline firms. The regularity of the pace of the corporate expansion process affects the success of overseas expansion. With a limited absorptive capacity, the return on investment obtained from rapid internationalization is obviously lower than that of gradual internationalization. For this paper, we selected the pace dimension of a firm’s internationalization process and classified this pace into regular and irregular cases to analyze the mechanism of the stability of the platform firm’s internationalization pace on its performance.

We propose the following hypothesis:

Hypothesis 1 (H1). 

The irregular internationalization pace of platform-based enterprises is negatively correlated with enterprise performance.

3.2. Moderating Role of the Network Effect of Platform Companies

Unlike traditional enterprises, platform-based enterprises are inherently located in a platform ecosystem. The ESA theory proposed by Li et al. [12] considers an ESA to be an FSA, which includes complementary resources, cooperation among platform firms, and rule-level advantages. Complementary resources and cooperation among platform firms reflect the indirect network effect of the platform. In the process of internationalization, platform companies need to break through their traditional boundaries and establish cooperation and strengthen coordination with companies and partners outside of these boundaries. The modularity and relationship-specific investments that platform companies often face cannot be explained by internationalization theory. On the one hand, modular architecture allows for collaborative innovation and production within the platform. Participants can determine if and when to join the platform by bundling their products with other complementary assets. On the other hand, participants must customize their products to the specifications of the platform. Internalization theory suggests that a modular business development model reduces the need for customization, while the need for relationship-specific transactions and investments reduces the need for modularity. The simultaneous existence of modularity and relationship-specific investments is self-contradictory. From an ecosystem perspective, platforms can implement policies that are open to the outside world, while modular management models can be used as tools for internal and external coordination. Thus, the internationalization of a platform ecosystem is based on relationship-specific investments made by participants, some of whom may join the ecosystem from the host country, while others need to customize their existing offerings to maintain a level of complementarity with local participants in the host country. This paper examines whether platform firms can use ESA as an FSA to moderate the adverse effects of rhythm instability on firm performance. The internationalization of platform firms has a network advantage over that of traditional firms. Using the Uppsala model to compare the internationalization of traditional firms and platform firms, it can be found that in the process of internationalization, platform firms preferentially use their own networks to achieve global expansion. The choice of the target market entry method is not the focus of platform firms’ business, as cooperation with local firms is more critical. Thus, we propose the following:

Hypothesis 2a (H2a). 

The platform network effect of platform-based firms weakens the negative effect of irregular internationalization rhythm on firm performance.

According to resource-based theory, a firm is a collection of resources [27]. The prerequisite for setting and achieving strategic goals is the availability of managerial, human, financial, and technological resources that ensure that a firm can conduct its normal business. When a company departs from its original business scope and develops more diversified business directions, it faces the risk and bottleneck of resource scarcity in the process of developing new markets, thus requiring more information input and support from external resources and knowledge for the company itself.

Without resources and knowledge, it is difficult for enterprises to effectively adapt to new business models during the process of entering a new market. From the perspective of resource acquisition and absorption, the information and knowledge reserves of enterprises are limited, and the internal accumulation process takes time, so it is difficult for enterprises to quickly reserve and transform information and knowledge in the unstable process of internationalization. In this process, enterprises are often unable to transfer information and knowledge to new businesses and often need to rely on the power of external relations. The social network of an enterprise can blur or even break through the boundary of exchange with the external society, making it possible to obtain resources and information from that society. For individual enterprises, access to resources and information is limited, and the social network relationships of enterprises can weaken the resource boundary of enterprises and expand their access to resources. Based on the above statement, we selected the widely recognized point degree centrality as a measure of social network node centrality from the perspective of the internationalization of platform-based enterprises and examined the moderating effect of social networks on the relationship between internationalization rhythm and the performance of these enterprises.

Hence, we propose the second sub-hypothesis of Hypothesis 2, as follows:

Hypothesis 2b (H2b). 

The social network effect of platform-based firms attenuates the negative effect of irregular internationalization rhythm on firm performance.

3.3. Moderating Role of Innovation Input of Platform Enterprises

In terms of R&D investment, Chinese enterprises that have already expanded internationally lack the experience and management ability to keep on growing. When entering the international market, these enterprises usually cope with the risks and disadvantages in the process of overseas expansion through technology and increased R&D, thus enhancing their innovation advantage. The irregularity of this international expansion trajectory accelerates investment in overseas R&D, thus reducing the negative impact of irregular overseas expansion. With the rapid development of internet digital technology, platform enterprises have become the main information resource medium in the new market. In the process of internationalization, the choice of competitive strategies by platform enterprises affects the degree of innovation and R&D investment. The sequence of internationalization of platform-based enterprises leads to different competitive environments. Early internationalizing firms chose to adapt to local policies, user habits, and the local environment to expand their overseas business scale. Platform-based enterprises that internationalized later faced the strong network base of the first entrants and needed to invest more in innovation and develop differentiated competition strategies to weaken their dominant market advantage. Platform-based enterprises can achieve disruptive innovation through their technological innovation advantages in market development and business model innovation by analyzing the competitive environment. The internationalization strategy of strengthening innovation investment can overcome the competitive disadvantage of a large number of existing platform companies due to homogeneity. With their innovative competitive advantages, platform companies can more easily enter and build user networks and complementary networks in emerging markets, thus weakening the uncertainties brought about by the irregular internationalization process. We consider platform firms’ innovation investment as a competitive advantage in their internationalization process, which can reduce the uncertainties brought about by an unstable internationalization rhythm and thus reduce its negative impact on performance.

Therefore, we propose Hypothesis 3, as follows:

Hypothesis 3 (H3). 

The innovation investment of platform-based firms weakens the negative effect of irregular internationalization rhythm on firm performance.

4. Materials and Methods

4.1. Sampling and Data

The sample was obtained from the data of China’s platform-listed companies, and the data sources were the financial statements from the official websites of major listed companies, such as the Guotaian database and the WIND database. Data on A-share-listed companies in Shanghai and Shenzhen from 2012 to 2020 were collected to investigate the impact of the internationalization rhythm of platform companies on corporate performance, of the moderating effect of the platform network effect, and of innovation investment on both.

With the focus of this study in mind, the sample was processed as follows. First, the business scope of A-share-listed companies in the Guotaian database was filtered by “platform”, “e-commerce”, “intermediary”, etc. The screened sample enterprises were checked against the information on the official websites of listed companies. After adjustment, 231 platform-type enterprises that met the requirements were identified. The abnormal enterprises marked as S and ST were excluded, and the panel data of 181 sample enterprises for 9 years from 2012 to 2020 were finally obtained.

4.2. Measurement of Variables

4.2.1. Explanatory Variables (Firm Performance)

The explained variable was the performance of platform-based firms, and the return on net assets, ROE (ROE = net income/average net assets × 100%) was chosen for measurement.

4.2.2. Explained Variables

The explanatory variable was the internationalization rhythm of platform enterprises. The rhythm of internationalization was first proposed by Vermeulen et al. [2]. It refers to the regularity and consistency of the international expansion mode of enterprises, and it is generally used to describe the degree of change in the internationalization expansion trajectory of enterprises, i.e., the change in the growth rate of overseas revenue of enterprises over a certain period of time or the degree of change in the establishment of overseas subsidiaries of enterprises over a certain period of time. This paper was based on the growth rate of the number of overseas subsidiaries. We measured the pace of internationalization based on the change in the growth rate of the number of overseas subsidiaries over a certain period of time. By referring to the empirical research of Zhang et al. [28] on the impact of corporate rhythm on performance, in this paper, a natural logarithm of the time variable (t) and overseas operating income (OI) within the annual window period was established [t − 4, t]:

$$Ln(O{I}_t)={\mathrm{b}}_1+{\mathrm{b}}_2\mathrm{t}+\mathsf{\delta },$$

The explanatory variable was defined as the natural logarithm of the firm’s overseas sales revenue in year t, and δ was the regression residual. We measured the firm’s internationalization rhythm by the inverse natural logarithm of the standard deviation of the regression coefficient b₂, as described by Chen et al. [29]. The larger the value is, the more inconsistent and regular the firm’s internationalization pace is.

In addition, to ensure the reliability of the findings, we further conducted robustness tests by referring to the studies of Wang et al. [30], Chang and Rhee [31], and Vermeulen and Barkema [2]. The internationalization pace of firms was measured based on the fluctuation of the number of their overseas subsidiaries over time, and the calculation formula is shown below. That is, the Kurtosis coefficient of the number of firms’ overseas subsidiaries during the annual window period [t − 4, t] was calculated, and the higher this value is, the stronger the irregularity of the firm’s internationalization expansion rhythm between year t and t + 4.

$$Rhythm=\left[\frac{n\left(n+1\right)}{\left(n-1\right)\left(n-2\right)\left(n-3\right)}\sum {\left(\frac{x_t-\overline{x}}{s}\right)}^4\right]-\frac{3{\left(n-1\right)}^2}{\left(n-2\right)\left(n-3\right)},$$

4.2.3. Moderating Variables

(1)

Social Network Effect

The most intuitive and widely recognized point centrality was selected as the measurement index to measure the node centrality of social networks. According to Zhang et al. [32], the higher the point degree centrality of a node in the network is, the more important it is compared with other nodes in the network. In undirected networks, point centrality is used to measure the degree of connection between a node and other nodes in the network. This paper adopted Freeman’s method and used degree centrality to reflect the centrality of nodes in the network in the social network [33]. Taking an undirected network with n nodes as an example, the point degree centrality of node i in the network is used to measure the total number of direct connections between i and the other n − 1 nodes.

The calculation formula is as follows:

$${\mathrm{Network}}_{\mathrm{i}}={{\displaystyle \sum }}_{\mathrm{j}-1}^{\mathrm{n}}\mathrm{a}\left(\mathrm{i},\mathrm{j}\right),$$

where Network denotes the point degree centrality of node i in the network, n is the total number of network nodes, and a(i,j) is a binary variable. a(i,j) = 1 when node i and node j are connected; otherwise, it is 0. The point degree centrality index of the social network was used to measure the richness of the social network of platform-listed companies [28]. We collected the concurrent appointments of directors and executives of platform-listed companies in other overseas companies, used the social network analysis software Ucinet 6.0 to transform the concurrent appointments of directors and executives of sample companies into a one-mode matrix, and then performed “Network-Centrality-Degree”. The point centrality of each node was calculated by the operation of “Network-Centrality-Degree”, and the point centrality index of the social network of platform-type listed companies from 2012 to 2020 was constructed.

(2)

Platform Network Effect

ESAs encompass complementary resources, platform–firm cooperation, and rule-level advantages. Considering data availability, we used the number of overseas affiliates of platform-listed companies to measure the platform network effect.

(3)

Enterprise Innovation Investment

Following Luo et al. [34], we used innovation investment intensity, which is measured by the ratio of R&D expenditure to operating revenue.

4.2.4. Control Variables

Drawing on Chow et al. [35] and Altaf et al. [36], we selected the following control variables: firm size (Size), firm age (Age), gearing (FL), board size (Board Scale), proportion of sole directors (Bodind), duality (Duality), and executive overseas background (Overseas Back). Year effect (Year) and industry effect (Ind) were also taken as control variables. In addition, all explanatory variables were lagged by one period to address possible endogeneity issues.

The Model and Methods Section includes the econometric testing of the model and variables and model setting. The selected variables were first tested for smoothness, and then the model form was tested and determined, keeping in mind the above-mentioned assumptions.

4.3. Modeling

To test the hypotheses, the following models were constructed. First, in Model 1, the independent and control variables were added, ${\mathsf{\beta }}_0$ was the intercept term and ε was the random disturbance term; in Model 2, the moderating variables Net, Social Net, and RD, and the regression coefficients were added; in Models 3, 4, and 5, the interaction terms of the moderating and independent variables were added, and the regression coefficients were ${\mathsf{\beta }}_4$ and ${\mathsf{\beta }}_5$, respectively, and Model 6 was tested for robustness by replacing the explanatory and explained variables.

Model 1: $\mathrm{ROE}$ = ${\mathsf{\beta }}_0$ + ${\mathsf{\beta }}_1$ Rhythm + ${\mathsf{\beta }}_{2\text{–}8}\sum \mathrm{controls}$ + $\mathsf{\epsilon }$.

Model 2:$\mathrm{ROE}$ = ${\mathsf{\beta }}_0$ + ${\mathsf{\beta }}_1$ Rhythm + ${\mathsf{\beta }}_2$ Net + ${\mathsf{\beta }}_3$ Social Net + ${\mathsf{\beta }}_4$ RD + ${\mathsf{\beta }}_{5\text{–}11}\sum \mathrm{controls}$ + $\mathsf{\epsilon }$.

Model 3: $\mathrm{ROE}$ = ${\mathsf{\beta }}_0$ + ${\mathsf{\beta }}_1$ Rhythm + ${\mathsf{\beta }}_2$ Net + ${\mathsf{\beta }}_3$ Social Net + ${\mathsf{\beta }}_4$ RD + ${\mathsf{\beta }}_5$ Rhythm × Net + ${\mathsf{\beta }}_{6\text{–}12}\sum \mathrm{controls}$ + $\mathsf{\epsilon }$.

Model 4:$\mathrm{ROE}$ = ${\mathsf{\beta }}_0$ + ${\mathsf{\beta }}_1$ Rhythm + ${\mathsf{\beta }}_2$ Net + ${\mathsf{\beta }}_3$ Social Net + ${\mathsf{\beta }}_4$ RD + ${\mathsf{\beta }}_5$ Rhythm × Social Net + ${\mathsf{\beta }}_{6\text{–}12}\sum \mathrm{controls}$ + $\mathsf{\epsilon }$.

Model 5: $\mathrm{ROE}$= ${\mathsf{\beta }}_0$ + ${\mathsf{\beta }}_1$ Rhythm + ${\mathsf{\beta }}_2$ Net + ${\mathsf{\beta }}_3$ Social Net + ${\mathsf{\beta }}_4$ RD + ${\mathsf{\beta }}_5$ Rhythm × RD + ${\mathsf{\beta }}_{6\text{–}12}\sum \mathrm{controls}$ + $\mathsf{\epsilon }$.

Model 6: $\mathrm{ROE}$ = ${\mathsf{\beta }}_0$ + ${\mathsf{\beta }}_1$ Rhythm + ${\mathsf{\beta }}_2$ Net + ${\mathsf{\beta }}_3$ Social Net + ${ \mathsf{\beta }}_4$ RD + ${\mathsf{\beta }}_5$ Rhythm × Net + ${\mathsf{\beta }}_6$ Rhythm × Social Net + ${\mathsf{\beta }}_7$ Rhythm × RD + ${\mathsf{\beta }}_{8\text{–}14}\sum \mathrm{controls}$ + $\mathsf{\epsilon }$.

5. Results

5.1. Test Results

This paper refers to the empirical research of Wang et al. [30] on the rhythm of corporate internationalization and used Stata 16.0 statistical analysis software to conduct descriptive statistics and correlation analysis on the model.

Table 1. Variable definition and sources.

	Variable Name	Variable Symbol	Measure Index	Variable Source
Explained variable	Platform enterprise performance	ROE	ROE = net income/average net worth × 100%	CSMAR
Explanatory variable	International rhythm	Rhythm	Kurtosis coefficient of the data within a time window of the number of overseas subsidiaries established by a firm in a given time period.	CSMAR/Manual collection of annual reports
Regulating variable	Platform network effect	Net	Number of overseas affiliates of platform companies	CSMAR/Manual collection of annual reports
	Social network effect	Social Net	Point degree centrality of social networks	CSMAR/Manual collection of annual reports
	Innovation input	R&D	R&D investment as a percentage of operating revenues	CSMAR
Control variable	Enterprise scale	InSize	Total corporate assets	CSMAR
	Enterprise age	Age	Length of time the company has been established is measured by the natural logarithm	CSMAR
	Asset-liability ratio	FL	Ratio of the enterprise’s total liabilities at the end of the period to total assets at the end of the period is measured	CSMAR
	Proportion of independent directors	Board Scale	Ratio of the number of sole directors to the total number of board members	CSMAR
	Two jobs in one	Bodind	Value of 1 is assigned when the chairman and the general manager are the same person; otherwise, it is 0	CSMAR
	Board Size	Duality	Total number of board members is measured by natural logarithm	CSMAR
	Overseas background of senior executives	Overseas Back	Ratio of the number of directors with overseas background to the total number of board members	CSMAR
	Dummy Variables	Year	Annual	CSMAR
		Ind	Industry	CSMAR

,

Table 2. Descriptive statistics and correlation coefficient matrix.

Variables	1	2		3	4	5	6	7	8	9	10	11	12
ROE	1
Rhythm1	0.031	1
Rhythm2	0.06	0.026		1
Net	−0.027	−0.026		−0.043	1
R&D	−0.022	−0.032		−0.022	0.022	1
lnSize	0.195 ***	0.006		−0.017	0.319 ***	−0.299 ***	1
Age	−0.045	−0.007		−0.011	0.119 ***	−0.344 ***	0.388 ***	1
FL	−0.143 ***	−0.065 *		0.024	0.125 ***	−0.469 ***	0.572 ***	0.242 ***	1
Board Scale	0.104 ***				−0.092 **	0.035	−0.136 ***	−0.137 ***	0.174 ***	0.042	0.160 ***	1
Bodind	−0.034				0.039	−0.006	0.114 ***	0.121 ***	−0.108 ***	−0.022	−0.102 ***	−0.410 ***		1
Duality	−0.043	0.018	−0.008	−0.045	0.117 ***			−0.205 ***	−0.190 ***	−0.150 ***	−0.238 ***	0.096 **	1
Overseas Back	0	−0.076 **	−0.032	0.073 *	0.092 **			0.022	−0.062	−0.022	0.144 ***	−0.026	0.087 **	1
Mean	5.229	3.25	2.132	6.3	7.143			22.679	17.56	40.358	8.6	37.946	0.314	0.993
Std. Dev.	15.473	5.1	0.806	9.729	9.065			1.228	5.709	19.751	1.708	6.111	0.464	1.056

Lower-triangular cells report Pearson’s correlation coefficients *** p < 0.01, ** p < 0.05, * p < 0.1.

,

Table 3. Regression results (dependent variable: corporate performance, ROE).

Variables	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Independent variable
Rhythm	−0.093 ***	−0.101 **	−0.094 *	−0.096 *	−0.082	−0.076 *
	(0.071)	(0.064)	(0.064)	(0.063)	(0.062)	(0.060)
Adjustment variables
Net		−0.126	−0.115	−0.125	−0.125	0.001
		(0.097)	(0.095)	(0.097)	(0.097)	(0.008)
Social Net		−0.321	−0.344	−0.357	−0.393	−0.001
		(0.878)	(0.880)	(0.904)	(0.868)	(0.001)
RD		−0.315 ***	−0.315 ***	−0.313 ***	−0.327 ***	−0.325 ***
		(0.099)	(0.099)	(0.099)	(0.095)	(0.094)
Moderating effects
Rhythm × Net			0.006 **			0.435 *
			(0.008)			(89.940)
Rhythm × Social Net				0.066 **		0.068 *
				(0.125)		(0.119)
Rhythm × R&D					0.0113 *	0.0112 *
					(0.006)	(0.006)
Control variables
lnSize	1.441	1.903 *	1.874 *	1.940 *	1.874 *	1.906 *
	(1.075)	(1.118)	(1.125)	(1.125)	(1.111)	(1.120)
Age	1.899	2.580 ***	2.611 ***	2.606 ***	2.836 ***	2.868 ***
	(1.195)	(0.709)	(0.701)	(0.705)	(0.628)	(0.633)
FL	−0.363 ***	−0.386 ***	−0.387 ***	−0.387 ***	−0.394 ***	−0.396 ***
	(0.059)	(0.056)	(0.056)	(0.056)	(0.056)	(0.057)
Board Scale	−0.568	−0.622	−0.618	−0.628	−0.631	−0.636
	(0.517)	(0.513)	(0.515)	(0.512)	(0.513)	(0.514)
Bodind	0.132	0.0965	0.102	0.095	0.11	0.109
	(0.081)	(0.086)	(0.088)	(0.086)	(0.085)	(0.088)
Duality	−0.862	−0.415	−0.386	−0.421	−0.241	−0.243
	(1.239)	(1.242)	(1.251)	(1.247)	(1.237)	(1.247)
Overseas Back	−0.618 *	−0.496	−0.505	−0.506	−0.516	−0.528
	(0.368)	(0.365)	(0.364)	(0.365)	(0.363)	(0.362)
Constant term	−3.353	−3.715	−3.767	−3.75	−4.023	−4.068
	(1.491)	(1.077)	(1.080)	(1.085)	(1.057)	(1.075)
R2	0.267	0.297	0.298	0.298	0.3	0.301
F value	5.18 ***	6.29 ***	6.06 ***	6.33 ***	7.01 ***	6.65 ***

Note: * is significant at the 10% level, ** is significant at the 5% level, *** is significant at the 1% level, and t values are in parentheses.

,

Table 4. Grouping robustness test regression results—Group 1.

Subgroup Robustness Test Group 1
Variables	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Independent variable
Rhythm	−0.096 **	−0.111 *	−0.0918 *	−0.105 **	−0.087	−0.0684 *
	(0.074)	(0.066)	(0.067)	(0.065)	(0.061)	(0.061)
Adjustment variables
Net		−0.145	−0.122	−0.145	−0.148	−0.132
		(0.099)	(0.098)	(0.099)	(0.100)	(0.098)
Social Net		−0.1317	−0.1382	−0.1335	−0.1415	−0.1476
		(0.101)	(0.102)	(0.103)	(0.099)	(0.103)
RD		−0.338 ***	−0.334 ***	−0.328 ***	−0.327 ***	−0.314 ***
		(0.123)	(0.123)	(0.119)	(0.120)	(0.116)
Moderating effects
Rhythm × Net			0.0109 **			0.00756 *
			(0.009)			(0.007)
Rhythm × Social Net				0.079 **		0.094 *
				(0.136)		(0.130)
Rhythm × R&D					0.010 ***	0.010 **
					(0.007)	(0.007)
Control variables
lnSize	1.078	1.465	1.43	1.532	1.497	1.551
	(1.133)	(1.139)	(1.132)	(1.169)	(1.125)	(1.152)
Age	0.459	0.508	0.725	0.606	1.561	1.769 *
	(1.168)	(0.920)	(0.797)	(0.848)	(0.984)	(0.958)
FL	−0.378 ***	−0.396 ***	−0.399 ***	−0.398 ***	−0.410 ***	−0.413 ***
	(0.069)	(0.067)	(0.068)	(0.067)	(0.069)	(0.070)
Board Scale	−0.141	0.0528	0.0822	0.0445	0.061	0.071
	(0.398)	(0.429)	(0.436)	(0.428)	(0.432)	(0.441)
Bodind	0.161	0.147	0.165	0.145	0.162 *	0.171 *
	(0.097)	(0.094)	(0.099)	(0.095)	(0.094)	(0.101)
Duality	−0.114	0.428	0.505	0.431	0.586	0.635
	(1.450)	(1.394)	(1.419)	(1.400)	(1.410)	(1.433)
Overseas Back	−0.399	−0.374	−0.393	−0.379	−0.409	−0.428
	(0.337)	(0.339)	(0.341)	(0.342)	(0.344)	(0.348)
Constant term	−12.55	−19.38	−22.48	−21.95	−34.29	−38.7
	(30.980)	(31.460)	(31.030)	(32.000)	(31.940)	(33.040)
R2	0.292	0.327	0.328	0.328	0.33	0.331
F value	3.44 ***	4.76 ***	4.87 ***	4.71 ***	5.32 ***	5.80 ***

Note: * is significant at the 10% level, ** is significant at the 5% level, *** is significant at the 1% level.

and

Table 5. Grouping robustness test regression results—Group 2.

Subgroup Robustness Test Group 2
Variables	Model 7	Model 8	Model 9	Model 10	Model 11	Model 12
Independent variable
Rhythm	−0.112 **	−0.136 *	−0.14 **	0.134	0.133	−0.131 *
	(0.309)	(0.292)	(0.291)	(0.294)	(0.323)	(0.320)
Adjustment variables
Net		−0.113	−0.114	−0.119	−0.113	−0.121
		(0.351)	(0.354)	(0.350)	(0.350)	(0.351)
Social Net		0.1659	0.1653	0.1767	0.1657	0.1755
		(0.201)	(0.201)	(0.194)	(0.203)	(0.196)
RD		−0.250 *	−0.250 *	−0.251 *	−0.252 *	−0.255 *
		(0.147)	(0.147)	(0.146)	(0.146)	(0.146)
Moderating effects
Rhythm × Net			0.007 *			0.008 *
			(0.052)			(0.056)
Rhythm × Social Net				0.086 **		0.086 *
				(0.274)		(28.940)
Rhythm × R&D					0.001 **	0.003 *
					(0.043)	(0.045)
Control variables
lnSize	3.22	3.566	3.538	3.55	3.564	3.517
	(2.734)	(3.252)	(3.253)	(3.259)	(3.262)	(3.270)
Age	11.17 ***	11.03 ***	11.06 ***	10.98 ***	11.03 ***	11.01 ***
	(2.355)	(2.346)	(2.358)	(2.356)	(2.353)	(2.381)
FL	−0.260 **	−0.330 **	−0.328 **	−0.329 **	−0.330 **	−0.327 **
	(0.124)	(0.131)	(0.132)	(0.133)	(0.131)	(0.135)
Board Scale	−1.061	−1.236	−1.242	−1.243	−1.238	−1.253
	(1.297)	(1.144)	(1.152)	(1.144)	(1.136)	(1.143)
Bodind	0.0207	0.0294	0.0266	0.032	0.0301	0.0311
	(0.147)	(0.183)	(0.184)	(0.185)	(0.181)	(0.185)
Duality	−3.441	−2.742	−2.761	−2.688	−2.729	−2.674
	(2.840)	(3.011)	(2.969)	(2.998)	(2.911)	(2.876)
Overseas Back	−2.830 **	−2.196	−2.21	−2.163	−2.188	−2.157 *
	(1.369)	(1.334)	(1.316)	(1.292)	(1.301)	(1.241)
Constant term	−174.0 ***	−175.9 ***	−175.5 ***	−175.2 ***	−175.9 ***	−174.7 ***
	(51.390)	(57.990)	(57.990)	(58.070)	(58.140)	(58.230)
R2	0.314	0.338	0.338	0.338	0.338	0.338
F value	4.12 ***	5.40 ***	3.88 ***	5.27 ***	5.94 ***	6.19 ***

Note: * is significant at the 10% level, ** is significant at the 5% level, *** is significant at the 1% level.

shows the definitions and sources of each variable. Table 2 reports the descriptive statistics of the main variables. As shown by the correlation coefficient matrix, the absolute values of the correlation coefficients among the explanatory variables were less than 0.5, which a priori indicates that there was no serious problem of multicollinearity. In addition, the results of the variance inflation factor (VIF) test indicated that the overall mean VIF was less than the threshold value of 2, and the VIF of all explanatory variables was much less than the threshold value of 10, which further confirmed that there was no multicollinearity problem. Table 3 shows the regression results.

In Table 2, it can be seen from the correlation coefficient matrix that the correlation of the main variables was not very significant. This is because of the high regularity and stability of platform companies, so their performance level will not be easily affected by fluctuations in the internationalization rhythm. This is also an aspect that is different from ordinary manufacturing companies.

In Table 3, Model 1 includes control variables and explanatory variables (Rhythm). The larger the inverse natural logarithm of the standard deviation of the regression coefficient of internationalization rhythm b₂, the lower the firm performance, i.e., the irregular internationalization rhythm of platform firms is negatively related to firm performance, which supports Hypothesis H1. Model 2 adds the platform network effect (Net), social network effect (Social Net), and firm innovation investment (R&D) to Model 1. Model 3 examines the moderating effect of the platform network effect on the relationship between the internationalization pace and the firm performance of platform firms. The results showed that the interaction term between the platform network effect and internationalization pace was significantly and positively related to firm performance ($\mathsf{\beta }$ = p < 0.05), and this result remained robust in the subsequent full Model 6. Model 4 examines the moderating effect of social network effects on the relationship between the internationalization pace and firm performance of platform firms. The results showed that the interaction term between the social network effect and the internationalization pace was significantly and positively related to firm performance ($\mathsf{\beta }$ = p < 0.05), and this result remained robust in the subsequent full Model 6. Model 5 examines the moderating effect of firm innovation investment on the relationship between the internationalization pace and the firm performance of platform firms. The test results showed that the interaction term between firm innovation investment and internationalization rhythm had a significant positive relationship with firm performance ($\mathsf{\beta }$ = p < 0.1), but the significance was not a strong network effect, and there was a competitive gap between successive entrants in the internationalization process of platform-based firms. Due to this difference, the first mover developed a stronger network advantage, and the innovation competitive advantage brought by the increase of innovation R&D investment advantage in effectively resisting uncertainty was not very significant, and this result remained robust in the subsequent full Model 6. Model 6 is a full-sample test with the addition of three interaction terms for the moderating variables platform network effect, social network effect, and firm innovation investment, and it can be seen that the correlation coefficient of the explanatory variable (Rhythm) was still negative and significant, and the correlation coefficients of all three interaction terms were positive and significant, further indicating that the above hypothesis was initially valid.

5.2. Robustness Test

To avoid the influence of linear regression on the significance and to ensure the robustness of the hypothesis testing results, we employed two types of robustness tests: grouping tests and alternative variables.

5.2.1. Grouping Method

From the structural theory perspective, not all individuals are closely connected in the social network of an enterprise, and the occupant of the structural hole position can seize the first opportunity to gain an information advantage and control advantage, and thus find opportunities to drive the enterprise to improve the performance. Additionally, due to the existence of network advantages of platform-based firms, there is a sequential entrant advantage in the order of establishing relationships with local partners in the process of overseas expansion. To determine the robustness of the hypothesis test in this paper, the study samples were grouped according to the criteria of internationalization before and after 2012, in which 84 samples were internationalized in 2012 and before (Group 1), and 97 samples were internationalized after 2012 (Group 2). The regression results were generally consistent and supported the above findings. The results of the grouping tests are shown in Table 4 and Table 5.

5.2.2. Alternative Variables

To ensure the reliability of the findings, we conducted further robustness tests by referring to Wang et al. [30], Chang and Rhee [31], and Vermeulen and Barkema [2]. The internationalization pace of firms was measured based on the fluctuation of the number of their overseas subsidiaries over time, and the calculation formula is shown below. That is, the Kurtosis coefficient of the number of firms’ overseas subsidiaries in the annual window period [t, t+4] was calculated, and the higher this value is, the stronger the irregularity of firms’ internationalization expansion rhythm between year t and t+4. The regression results were generally consistent and supported the above conclusions. The results of the subgroup tests are shown in

Table 6. Replacing the regression results of independent variables.

Robustness Tests for Alternative Variables
Variables	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Independent variable
Rhythm2	−0.533 **	−0.486 *	−0.483 **	−0.473	−0.47	−0.453 *
	(0.509)	(0.524)	(0.526)	(0.522)	(0.522)	(0.520)
Adjustment variables
Net		−0.118	−0.118	−0.12	−0.118	−0.121
		(0.101)	(0.101)	(0.100)	(0.100)	(0.100)
Social Net		−0.755	−0.757	−0.748	−0.745	−0.732
		(0.996)	(0.996)	(0.970)	(0.101)	(0.986)
R&D		−0.454 **	−0.454 **	−0.453 **	−0.475 ***	−0.476 ***
		(0.188)	(0.189)	(0.187)	(0.179)	(0.177)
Moderating effects
Rhythm2 × Net			0.00731 *			0.0024 *
			(0.033)			(0.033)
Rhythm2 × Social Net				85.26 *		91.83 *
				(49.320)		(50.620)
Rhythm2 × RD					0.0784 ***	0.0873 **
					(0.095)	(0.096)
Control variables
lnSize	1.226	1.5	1.494	1.532	1.424	1.449
	(1.286)	(1.267)	(1.271)	(1.269)	(1.273)	(1.281)
Age	1.901 *	−0.672	−0.665	−0.494	−1.766	−1.696
	(1.107)	(2.317)	(2.349)	(2.470)	(3.237)	(3.501)
FL	−0.395 ***	−0.412 ***	−0.412 ***	−0.408 ***	−0.411 ***	−0.407 ***
	(0.059)	(0.059)	(0.059)	(0.059)	(0.060)	(0.060)
Board Scale	−0.358	−0.33	−0.332	−0.315	−0.321	−0.305
	(0.580)	(0.588)	(0.590)	(0.592)	(0.596)	(0.603)
Bodind	0.16	0.114	0.114	0.112	0.12	0.118
	(0.097)	(0.092)	(0.092)	(0.093)	(0.091)	(0.092)
Duality	−0.836	−0.401	−0.394	−0.414	−0.417	−0.431
	(1.369)	(1.329)	(1.330)	(1.337)	(1.333)	(1.343)
Overseas Back	−0.625	−0.509	−0.507	−0.545	−0.486	−0.521
	(0.409)	(0.404)	(0.404)	(0.402)	(0.403)	(0.402)
Constant term	−33.37	−1.042	−1.005	−4.328	14.57	12.82
	(34.300)	(43.710)	(43.990)	(44.980)	(53.680)	(56.420)
R2	0.281	0.315	0.315	0.318	0.317	0.32
F value	4.91 ***	4.71 ***	4.48 ***	5.88 ***	4.52 ***	65.39 ***

Note: * is significant at the 10% level, ** is significant at the 5% level, *** is significant at the 1% level.

.

6. Discussion

In this paper, a multiple linear regression model was used to test the hypothesis, and the results of each regression are shown in

Table 7. Regression statistics of empirical results.

Assumption No.	Hypothetical Content	Empirical Results
H1	The irregular internationalization rhythm of platform firms is negatively related to firm performance	Support
H2a	The platform network effects of platform firms weaken the negative impact of irregular internationalization rhythm on firm performance	Support
H2b	The social network effects of platform firms weaken the negative effects of irregular internationalization rhythms on firm performance	Support
H3	The innovation investment of platform firms weakens the negative effect of irregular internationalization rhythm on firm performance	Support

.

Internationalization is a dynamic process by which its pace is an important indicator of success. The rhythmic internationalization process enhances the sustainable development capabilities of platform-based enterprises and ultimately improves their performance. When the internationalization rhythm is unstable, some companies may face additional risks. In fact, the internationalization rhythm of platform companies and the internationalization rhythm of manufacturing companies have different impacts on their firm performance. Unlike platform companies, most manufacturing companies are relatively small and lack resources and capabilities. Especially for some manufacturing companies lacking technological innovation and knowledge integration capabilities, their performance levels will be more affected by many factors, such as internationalization rhythm patterns, corporate network locations, and dynamic capabilities [37]. Therefore, the internationalization rhythm and performance of manufacturing companies have a relatively high correlation. In addition, when a manufacturing company is in the early stage of international expansion, it will face the challenges of changing overseas market environments, and manpower, equipment allocation, internal management, and external business brought by operating in a new market. This will increase the cost of the company (LON) [38]. Manufacturing companies also face Liability of Foreignness (LOF), including financial risk for operating in overseas markets. Due to the Liability of Newness and Liability of Foreignness, the internationalization rhythm of manufacturing companies will inhibit the performance of companies. Therefore, the internationalization rhythm of manufacturing companies is negatively correlated with firm performance. However, for large-scale platform companies, the amount of capital and management capabilities they possess are significantly higher than those of manufacturing enterprises. Platform companies not only have more innovation resources to invest in overseas expansion, but also often have a stronger ability to digest and absorb innovation resources. At the same time, platform companies have strong resource stickiness, which can help companies maintain high regularity, stability, and coherence in international development [39]. This means that even if the international rhythm of platform companies is disrupted, they can still maintain stable development. Platform companies are more flexible in terms of international transactions and innovation resource integration in the international rhythm [40]. Therefore, the performance of platform companies will not be greatly negatively affected by the instability of the internationalization rhythm. In addition, the stable and regular international expansion strategy of platform companies will more effectively build overseas platform networks. The rhythmic internationalization process enhances the sustainable development capabilities of platform companies and ultimately improves the company’s performance [41]. In conclusion, the regular internationalization rhythm of platform companies has a positive impact on company performance, but its correlation is weaker than that of manufacturing companies.

The network effect of platform-based enterprises can significantly moderate the relationship between their internationalization rhythm and their overall performance. In particular, the number of enterprises’ overseas affiliates can reflect the network size and network diversity of platform-based enterprises in host countries, and stronger overseas enterprise cooperation enhances the indirect network effect of the platform and helps enterprises strengthen a systematic coordination with partners beyond the organizational scope. In platform ecosystems built by firms and partners, a distinguishing feature of platform enterprises is the malleability of digital innovation, which allows for platform design, governance rules, and ecosystem scope to be adjusted after platform launch. This makes experimental learning particularly valuable and allows platform-based companies to identify the best ways to improve complementarities within the ecosystem as they evolve. The biggest difference in the internationalization risk for platform-based firms compared to traditional firms is in terms of commercial risk. Since the core services provided by a platform do not significantly change when entering different target markets, firms need to rely on local partners who know the target market better to provide the physical assets needed to operate. Partners have the prerogative to choose the partner platform and capital investment. Therefore, the choice of local partner plays a crucial role in the development of the platform in the target market.

Regarding the moderating effect of the social network effect of enterprises, we selected the point degree centrality index of enterprise social networks to measure the richness of enterprise social networks, and we verified that the richness of platform enterprise social networks can effectively reduce performance reduction due to irregular internationalization proceedings. Diversified and rich social network resources can effectively help directors, executives, and managers face the risk problems brought by information asymmetry in a shorter period of time, and enterprises’ demand for information through social network channels is more urgent in the context of the platform economy. These resources can help enterprises achieve network connectivity through a higher digital access level so that they can make even more responses and adjustments in platform strategy and rules. At the same time, a rich social network allows decision-makers to reduce the cost of trial and error in unfamiliar environments and better learn from the experience of local partner enterprises so that their firms can achieve better performance levels.

The moderating effect of corporate innovation R&D investment is significant, as shown in the empirical part of the regression results. Platform-based enterprises with sustainable development capabilities can break through the boundary constraints of innovation elements’ strengthening, thus enhancing their competitive advantage, and improving their innovation performance [42]. When group robustness tests were conducted on the sample of platform firms, it can be found that for platform firms that internationalized at the early stage, innovation investment more significantly moderated the adverse effects of irregular internationalization on performance; however, for platform firms that internationalized at the later stage, the moderating effect of innovation investment significantly decreased.

With the popularization of the Internet sites and AI technology, platform-based enterprises have gradually become the main medium of electronic payments in emerging markets, and platform-based enterprises that undergo internationalization at a later stage face higher competitive pressures from the host country. Rongshuai et al. [43] proposed that the user installation base and product network externalities owned by incumbent firms are still important obstacles that platform firms need to overcome in their cross-border operations. The stronger the product network externality is, the greater the degree of product compatibility that platform firms adopt to consider consumers’ consumption habits, thus reducing the degree of cross-border innovation. Unlike traditional latecomer enterprises that take disruptive innovation as their main market strategy, platform enterprises choose technological innovation breakthroughs in addition to market innovation breakthroughs, and the more obvious their user resource advantages and cross-market network effects are, the more innovative their cross-border products are. When facing competitors who build ecological networks locally at an early stage, the internationalization strategy of latecomer firms in strengthening innovation investment can overcome the competitive disadvantage of a large number of existing platform firms due to homogenization. Nevertheless, the network effect among user networks and complementary networks built by competing firms at an early stage is stronger, and the competitive advantage brought by innovation investment will be relatively small. Thus, the significance of the moderating effect on the reduced performance of firms due to irregular internationalization is also reduced.

7. Conclusions

In conclusion, for Chinese platform companies, internationalization is an important development direction and a key to their future success. By making full use of the social network effect and developing practical internationalization strategies, platform companies can improve their international competitiveness and performance, and thus grow faster in the international market.

In addition, the study of the pace of internationalization, social network effects, and performance of Chinese platform companies has a positive effect on a similar global research environment and has contributed to the advancement of platform companies globally in several ways: (1) Chinese platform companies have grown significantly in the past few years. The study of their internationalization pace, social network effects, and performance has opened up new areas of research for academics. An in-depth study of the cases and experiences of Chinese platform-based companies can expand the theoretical understanding and practical application of platform-based firms and extend this knowledge to similar research settings, globally. (2) Chinese platform-based companies have generated a wealth of data during their internationalization processes, which can provide support for empirical analysis in similar research settings, globally. By analyzing this data, the pace of internationalization, social network effects, and performance can be verified and measured for applicability in other regions and further promote research and practice on a global scale. (3) Chinese platform companies have accumulated valuable experience and lessons learned in the process of internationalization. These experiences have implications for companies in similar research environments around the world. By studying the success and failure cases of Chinese platform-based companies, useful lessons can be provided to companies in other countries or regions to help them better cope with challenges and improve their performance in the internationalization process. As such, research on the international development of platform-based companies provides a useful resource for theoretical guidance, knowledge expansion, data support, and empirical analysis, as well as lessons learned, to promote platform-based companies on a global scale.

Due to time and cost constraints, the sample for this study was limited to the period 2010 to 2020. The empirical results of this study do not provide a complete picture of the characteristics of the internationalization process of Chinese companies over the long term. In addition, there are other variables not considered in the model of this study, such as changes in the competitive landscape or macroeconomic conditions, which may also affect the performance of firms. Therefore, there are a number of future research directions that deserve the attention of scholars. Firstly, future research could explore other variables that may moderate the relationship between internationalization rhythms and firm performance to better understand the factors that affect firm performance in a global context. Secondly, social network effects have a significant impact on the internationalization process of platform-based firms. Future research could explore how platform firms use social networks to expand their international partner networks, attract global users and ecosystem players, and enable value co-creation and innovation through social networks. Finally, studying the pace of internationalization of Chinese platform companies is crucial to understanding their global expansion strategies and international competitiveness. Future research could focus on different industries and types of companies and delve into their internationalization strategy choices, overseas market entry strategies, as well as their operating models and performance in international markets.