The Influence of Web 2.0 Tools on the Sustainable Development of E-Commerce: Empirical Evidence from European Union Countries

Abstract

In the context of accelerating digitalization, this study investigates how electronic commerce performance is influenced by Web 2.0 instruments in the 27 EU member states. Analyzing literature reviews and performing our own bibliometric review, we identified a gap related to the measurable economic results of e-commerce. The scope of this study was to analyze the relationship between Web 2.0 tools and the level of turnover generated by e-commerce, applying robust econometric models based on panel data regression with random effects and fixed effects (Arellano–Bond). The results highlight that the online paid advertisement and social media usage variables have significant, positive effects on e-commerce performance, confirming the first and second hypotheses. “Use the enterprise’s blog or microblogs” and “use of multimedia content sharing websites” do not influence enterprises’ total turnover from e-commerce sales to a valid and statistically significant extent. Thus, the third and fourth hypotheses are not confirmed by the results of the research conducted, possibly due to limited innovation and platform ownership in Europe. This study makes a notable empirical and methodological contribution, embedding digital sustainability in the analysis, which implies that the findings can be used for updating e-commerce policies.

1. Introduction

Digital transformation significantly influences enterprises and the way they conduct their activities, especially regarding electronic commerce [1], known as e-commerce. One of the essential factors for this evolution is the extensive usage of Web 2.0 instruments [2], which facilitates collaboration, co-creation, and interactions with consumers [3]. The economic environment, organizations, and consumers behavior [4] have been strongly impacted by the power of digital transformation. The digitalization of the economy at the level of the European Union states is a priority, as shown through diverse initiatives and programs [5], integrating Web 2.0 instruments in digital transformation and sustainable development policies. Consequently, the analysis of how e-commerce development is influenced by Web 2.0 tools is mandatory in the context of rapid digitalization due to the pandemic and economic sustainability.

The role of social media and online advertisement [6] in influencing customer behavior has been extensively studied in the literature, but the concrete economic outcome of utilizing these instruments has been insufficiently studied for European countries. The scientific literature from the field is concentrated on themes related to social effects and comportments, but the subject of economic performance has not been explored in a quantitative and comparative way. Taking into consideration that the results of these technologies have not been sufficiently studied from an economic point of view, we plan to study how Web 2.0 tool adoption influences e-commerce performance in the EU member states.

Our bibliometric review confirms the presence of a gap in quantitative analysis, because the majority of the works are qualitative, interview, or case study-based, focusing on a single company, industry, or country. In addition, the focus until now has headed towards consumer behavior and towards technology advancements, without giving enough attention to the measurable economic effects generated by the relationship between Web 2.0 instruments and e-commerce. This study aims to contribute to the literature, through applying econometric models on panel data gathered from the 27 EU member states, for the 2015–2023 timeframe.

The structure of this article is as follows: We conduct a critical review of the scientific literature with a focus on Web 2.0, e-commerce, economic sustainability, a bibliometric review, identifying gaps in the literature, and developing hypotheses. We design a research methodology based on a quantitative approach of panel data regression based on the Eurostat database. We present the research results with a focus on coefficients for fixed-effects and random-effects variables; a random-effects estimation summary; random effects—parameter estimates; fixed effects—parameter estimates; robust standard errors, clustered by both country and year; robust standard errors—parameter estimates; and a dynamic panel model (Arellano–Bond style)—parameter estimates. We also present a discussion with a focus on novelty findings, as well as conclude this paper. The scope of this study is to analyze the relationship between Web 2.0 tools and the level of turnover generated by e-commerce, applying robust econometric models based on panel data regression with random effects and with fixed effects (Arellano–Bond).

The link between Web 2.0 instruments and the level of turnover performed by e-commerce is analyzed through applying strong econometric models based on panel data regression with random effects and fixed effects (Arellano–Bond). This study puts emphasis on the economical sustainability dimension and focuses on specific Web 2.0 tools which are individually evaluated. The results underscore that the online paid advertisement and social media usage variables have a significant, positive effect on e-commerce performance, while “use the enterprise’s blog or microblogs” and “use of the multimedia content sharing websites” do not influence enterprises’ total turnover from e-commerce sales to a valid and statistically significant extent. Because data are gathered from 27 EU member states, the conclusions show common patterns at the European level.

2. Materials and Methods

2.1. Analyzing the Scientific Literature Related to Web 2.0 and E-Commerce

Web 2.0 is the second generation of internet-based services, with a focus on collaboration, social media, and content generated by users [7]. It has various instruments such as social media platforms, blogs, vlogs, and recensions [8]. Its tools play an important role in building the link between clients and companies, having a large influence on consumer behavior. E-commerce development is usually measured by indicators such as the number of companies involved in e-commerce activities, percentage of digital transactions, and online sales turnover [4]. Web 2.0 technology allows collaboration, two-way interaction, and, finally, user content creation, which has a large impact on the digital transformation of the global economy in general. Through this, marketing strategies and organizational architecture are also part of the remodeling process [9]. Web 2.0 tools are more than promoting channels; they have the potential to innovate and collect feedback and collaborative content, in order to answer to the constantly changing needs of the clients. SMEs have become increasingly more open to utilizing Web 2.0 tools, but the level of adoption differs depending on organizational culture, available resources, and sector activity [10]. Communication with clients, growth in visibility of the brand, and number of online purchases were improved through the adoption of Web 2.0 tools, which has the power to influence shopping behavior [2].

The usage of Web 2.0 has made a strong contribution to economic sustainability [11] because it reduces the access barriers to markets, it reduces the costs through digital marketing, and it promotes responsible consumption [12]. Web 2.0 is associated with the circular economy [13], responsible consumption, collaborative business [14], going beyond communication, and facilitating economic sustainability at all levels [15]. The sustainability of firms is optimized at economic, ecological, social, and governmental levels through the use of Web 2.0 instruments, evolving into transparency. These firms are more adaptable to changes and crisis because they are able to optimally coordinate their resources [16] and to efficiently communicate with their stakeholders, building their sustainable resilience [17]. Through these tools, the ecological footprint is indirectly reduced, because the costs associated with traditional marketing are decreased [18].

The European Union is aiming to harmonize digital capabilities [19] in order to support e-commerce participation in the member states. There is a significant difference between the member states considering the level of e-commerce adoption [20] and digital infrastructure. The relationship between Web 2.0 tools and the subject of e-commerce performance was studied in the scientific literature by various scientists. Tajvidi and Karami (2021) evaluated the social media impact on the performance of EU companies through structural equation modeling [21]. The literature shows that there exist a positive link between the performance of e-commerce and Web 2.0 instruments [22]; Ahmad et al., 2019 concluded in their articles that the companies from Europe and Asia with an online presence in social media will have superior performance in terms of sales numbers compared to companies which use only traditional methods [23]. The outcomes depend on the national context; digital development level in general; and accessibility, governance, and types of concurrence. Consequently, the economic impact is not always seen in the short term. In this regard, recent studies advocate for the use of robust quantitative models [24], such as panel data regressions and dynamic models, which can isolate the specific effects on financial performance [25]. In the scientific literature, Web 2.0 has been analyzed over the years through different theoretical frameworks, such as the Technology Acceptance Model (TAM) [26], Diffusion of Innovation Theory [27], and Resource-Based View (RBV) [28]. The first two frameworks explain that the adoption of new technologies can be influenced by the perception of the users related to ease of use [29].

In order to better understand the relationship between Web 2.0 and e-commerce, a bibliometric review was performed. The search was performed in the Web of Science Database with the keywords Web 2.0 AND e-commerce, for the period 2015–2025 and for the document types of articles and review articles, resulting in 209 documents. The analysis tools used in this case were VosViewer version 1.6.20 and Biblioshiny from Bibliometrix 5.0 R package. The first two research areas in this field are business economics and computer science, with the most prolific countries in this regard being China, Australia, and England [30]. The keyword map contains 3 clusters, 33 items, 371 links, and a total link strength of 887 [31]. The “commerce” keyword is central, having the highest number of connections. The first cluster with the keywords commerce, consumer, platform, internet, and social commerce suggest consumer behavior after technology adoption. The second cluster with the keywords information, analysis, adoption, opportunity, role, and website focuses on the technological part of e-commerce. The third cluster with the keywords product, service, and time can represent the general methodological part of the research. The map underlines the high level of interdisciplinarity, with social, technological, and methodological themes. Starting with 2015 and running until 2021, there was an increasing trend of publications, with a peak in 2021; after the highest registered score, the number of publications declined, with a minimum registered in 2024, but in 2025, there was a slight rebound. The citation followed the trend of publications between 2014 and 2021, with a citation peak in 2021; after 2021, there was a slight decrease. The 2021 year was a prolific year for both publications and citations, having a major academic impact, which can be explained by the interest in the field due to the pandemic. After 2021, interest in this subject dropped, implying a possible saturation or change in interest of the academic community.

The most influent articles, shown in

Table 1. Most influent articles.

Article	Subject of the Article	Number of Citations
Social commerce constructs and consumer’s intention to buy	E-commerce and consumers’ intention to buy online through an SEM quantitative methodology [2]	491
Understanding social commerce: A systematic literature review and directions for further research	Systematic literature review of social commerce [32]	247
Social commerce research: Definition, research themes, and the trends	User-centric research in the field of e-commerce [1]	185
The effects of Facebook browsing and usage intensity on impulse purchase in f-commerce	Data impulse purchases in the field of e-commerce [33]	139

Source: Authors using Biblioshiny from Biliometrix 5.0 R package.

, with the most registered number of citations, treat the following subjects:

After analyzing the literature, we found gaps related to the lack of usage of multiple Web 2.0 instruments in explaining e-commerce. In addition, only a few articles use comparative studies between countries based on data panels from the EU. There is a lack of dynamic models and quantitative studies, with the majority of articles focusing on individual cases and qualitative aspects. The e-commerce and Web 2.0 link is a studied topic in the scientific literature, but there are various limitations which can be studied further. On one hand, a significant part of the papers concentrate on qualitative analysis [2], exploring the perceptions of the users and digital marketing strategies [34]. On the other hand, the quantitative analysis on the subject focuses on a single state or specific industry [23], without an international comparison background [21]. As a result, digital advantages can be seen in the long term and the relationship between Web 2.0 and e-commerce cannot be fully explained by static models.

E-commerce is evolving to become a key factor of sustainable competitiveness in the European Union. Taking into consideration the limitations that exist in the literature, there is a need for advanced econometrical models, such as dynamic models based on panel data, for example, Arellano–Bond models, which can explain the temporal effects of the usage of Web 2.0 tools on commercial performance.

2.2. Development of Hypotheses

In the digital age, consumer behavior is complex, with companies making countless efforts to attract and retain consumers in virtual spaces and to maintain their interest in purchasing goods and services [35]. The clients in online channels are sometimes generated through paid advertising in online platforms [36]. Previous studies have highlighted the positive relationship between Internet advertising features like search advertisement and classified advertisement and e-commerce sales in the case of five European countries [37]. Therefore, we propose the following hypothesis:

H1. 

Pay to advertise on the internet positively influences the enterprises’ total turnover from e-commerce sales.

In the Web 2.0 context, most companies, including SMEs, use social networks (e.g., Facebook and LinkedIn) to open new opportunities in customer relationships and to redefine business processes and maximize their e-commerce value. Social networks like Facebook, Instagram, Twitter, and LinkedIn provide companies with the ability to directly engage with consumers, building brand awareness and fostering customer loyalty [9]. Research has shown that these platforms allow companies to reach a broader audience and engage in real-time conversations, which is critical in driving e-commerce turnover: social media serves as both as a communication tool and a marketing channel, influencing consumer decisions in multiple stages [8]. Integrating social media marketing efforts can increase both customer engagement and overall e-commerce sales, as clients are increasingly expecting seamless experiences across all platforms [38]. For these reasons, we propose the following hypothesis:

H2. 

Use social networks (e.g., Facebook and LinkedIn) positively influences the enterprises’ total turnover from e-commerce sales.

The frequency and the type of post on microblogs significantly impacted consumer engagement, which in turn increased sales [39,40]. Freberg et al. (2011) highlighted that influencer endorsements on platforms like Twitter or LinkedIn often result in increased consumer trust, which directly correlates with increased sales [41]. Microblogging allows for the rapid sharing of content that can reach millions of users, often resulting in viral promotions that drive traffic to e-commerce platforms. Influencer-generated content and brand mentions can lead to higher engagement and increased e-commerce turnover, as influencers tend to attract highly engaged followers who trust their recommendations [42]. Therefore, we propose the following hypothesis:

H3. 

Use the enterprise’s blog or microblogs (e.g., Twitter and Present.ly) positively influences the enterprises’ total turnover from e-commerce sales.

Multimedia content sharing websites, such as YouTube, Flickr, and Vimeo, have become central to digital marketing strategies due to their ability to drive engagement, build brand awareness, and directly influence e-commerce turnover. Incorporating influencers into multimedia campaigns on platforms, like YouTube, can amplify brand exposure, especially when influencers create content that aligns with the brand’s identity and messaging. This has been shown to enhance brand recall and purchase intent, which leads to increased e-commerce revenue [43].

H4. 

Use multimedia content sharing websites (e.g., YouTube and Flickr) positively influences the enterprises’ total turnover from e-commerce sales.

3. Research Methodology

A quantitative approach based on panel data regression was followed for this research. The data sources used were Eurostat databases on the following: enterprises’ total turnover from e-commerce sales (dataset: value of e-commerce sales by size class of enterprise; unit of measure: percentage of turnover) as an expression of e-commerce development, pay to advertise on the internet, use of social networks (e.g., Facebook, LinkedIn, Xing, Viadeo, and Yammer), use the enterprise’s blog or microblogs (e.g., Twitter and Present.ly), and use multimedia content sharing websites (e.g., YouTube, Flickr, Picasa, and SlideShare; dataset: social media use by type; internet advertising and size class of enterprise; unit of measure: percentage of enterprises).

Data from the period of 2015–2023 were selected for the dependent variable and the four independent variables for the 27 countries of the European Union. The period was chosen by taking into account the years in which all data were available for all five variables of the model. In addition, the 27 countries represent the majority of the most advanced and representative in Europe both in terms of the development of Web 2.0 and its tools as well as digital infrastructure and e-commerce. The time interval chosen for the panel data regression is particularly relevant for the purpose of the research conducted because it represents the maturity period of Web 2.0 tools and their potential effects on e-commerce development.

Table 2. Variables and data sources.

No.	Variable	Variable Type	Data Source	Dataset
1	Enterprises’ total turnover from e-commerce sales	Dependent	Eurostat	Value of e-commerce sales by size class of enterprise; unit of measure: percentage of turnover
2	Pay to advertise on the internet	Independent	Eurostat	Social media use by type, internet advertising and size class of enterprise; unit of measure: percentage of turnover
3	Use social networks (e.g., Facebook, LinkedIn, Xing, Viadeo, and Yammer); use the enterprise’s blog or microblogs (e.g., Twitter and Present.ly);	Independent	Eurostat	Social media use by type, internet advertising and size class of enterprise; unit of measure: percentage of enterprises
3	use the enterprise’s blog or microblogs (e.g., Twitter and Present.ly)	Independent	Eurostat	Social media use by type, internet advertising and size class of enterprise; unit of measure: percentage of enterprises
4	Use multimedia content sharing websites (e.g., YouTube, Flickr, Picasa, and SlideShare).	Independent	Eurostat	Social media use by type, internet advertising and size class of enterprise; unit of measure: percentage of enterprises

Source: Authors.

presents the details on the variables and datasets.

In order to choose the most appropriate regression model between the fixed-effects model estimated with entity and time effects and the random-effects model, the Hausman test was applied [44]. The results obtained from the Hausman test showed that the random-effects model is preferred, more efficient, and statistically justified. However, we also preferred to apply the fixed-effects model given that countries likely have unobserved characteristics affecting e-commerce. In both the random-effects and fixed-effects models, enterprises’ total turnover from e-commerce sales was considered the dependent variable, and pay to advertise on the internet, use of social networks (e.g., Facebook, LinkedIn, Xing, Viadeo, and Yammer), use the enterprise’s blog or microblogs (e.g., Twitter and Present.ly), and use multimedia content sharing websites (e.g., YouTube, Flickr, Picasa, and SlideShare) were considered independent variables.

Although the same variables (pay to advertise on the internet, use of social networks) were obtained in the fixed-effects model as in the random-effects model, being those that positively influence enterprises’ total turnover from e-commerce sales, several statistical tests were applied for the fixed-effects regression model: the Breusch–Pagan test for heteroskedasticity [45]; the Woolridge test for serial correlation [46]; and the Pesaran Test for cross-sectional dependence [47]. Since the Woolridge test indicated that a serial correlation is present in the panel data residuals and the Pesaran Test showed that there is a cross-sectional dependence, a re-estimated fixed-effects model using robust standard errors, clustered by both country and year, was first re-estimated. For the serial correlation problem, we estimated a dynamic panel model (Arellano–Bond style), adding a lagged dependent variable.

4. Research Results

To confirm the research hypotheses and to quantify the positive influence of Web 2.0 tools on the essential elements that illustrate the development of e-commerce (such as enterprises’ total turnover from e-commerce sales), panel data regression was used. To choose the regression model between the fixed-effects model estimated with entity and time effects and the random-effects model, the highly relevant Hausman test was applied, as it shows which estimation method is the most appropriate. The null hypothesis in the case of the Hausmann Test is as follows: the random-effects estimator is consistent and efficient (i.e., unobserved heterogeneity is uncorrelated with regressors). The alternative hypothesis is as follows: only the fixed-effects estimator is consistent (i.e., unobserved heterogeneity is correlated with regressors). Hausman test coefficients for fixed-effects and random-effects variables are presented in

Table 3. Coefficients for Fixed Effects and Random Effects variables.

Variable	Fixed Effects	Random Effects	Difference
Pay to advertise on the internet;	0.1125	0.2754	−0.1629
Use social networks (e.g., Facebook, LinkedIn, Xing, Viadeo, and Yammer);	0.0060	0.0793	−0.0733
Use the enterprise’s blog or microblogs (e.g., Twitter and Present.ly);	0.0772	0.1177	−0.0406
Use multimedia content sharing websites (e.g., YouTube, Flickr, Picasa, and SlideShare).	−0.0077	0.0174	−0.0251

Source: own calculations.

.

Hausman test outputs (Hausman statistic = −664.7443; degrees of freedom = 4; p-value = 1.000) fail to reject the null hypothesis at the 5% level and suggest that the random-effects model is preferred (more efficient) and statistically justified. The random-effects model could be used, is more efficient, but assumes no correlation with unobserved heterogeneity.

The random-effects model, from Figure 1, estimates the impact of various online activities on e-commerce turnover across countries and years. The dependent variable is e-commerce turnover and the independent variables are as follows: pay to advertise on the internet; use social networks; use the enterprise’s blog or microblogs; and use multimedia content sharing websites. The random-effects estimation summary is presented in the figure below. The model explains 83.3% of the variation between countries because R2 (between) is 0.8326. The value of R2 (within), 0.1108, shows that the model explains 11.1% of the variation within countries over time. The F-statistic value (38.004; p < 0.001) indicates that the model is highly significant overall.

Random-effects parameter estimates are presented in

Table 4. Random Effects—Parameter Estimates.

	Coefficient	Std.Err	T-Stat	p-Value	Lower CI	Upper CI
Pay to advertise on the internet;	0.2754	0.0398	6.9135	0.0000	0.1969	0.3538
Use social networks (e.g., Facebook, LinkedIn, Xing, Viadeo, and Yammer);	0.0793	0.0248	3.1938	0.0016	0.0304	0.1283
Use the enterprise’s blog or microblogs (e.g., Twitter and Present.ly);	0.1177	0.1248	0.9435	0.3464	−0.1281	0.3635
Data Use multimedia content sharing websites (e.g., YouTube, Flickr, Picasa, and SlideShare).	0.0174	0.0111	1.5670	0.1184	−0.0045	0.0392

Source: own calculations.

. The most influential factor and the strongest predictor in the model are the percentage of enterprises paying to advertise online (β = 0.2754; p < 0.001). For every 1 percentage point increase in online advertising, e-commerce turnover rises by about 0.28 percentage points.

Social network usage (β = 0.0793; p = 0.0016) is also significant at the 1% level: a 1 percentage point increase in enterprise use of social networks is associated with a 0.08 percentage point increase in e-commerce turnover. The use of the enterprise’s blog or microblogs (β = 0.1177; p = 0.3464) and multimedia content sharing websites (β = 0.0174; p = 0.1184) do not have a statistically significant effect on e-commerce turnover in this model.

Although the Hausman test statistically favors the random-effects (RE) model, we also considered the fixed-effects (FE) specification, following both theoretical and empirical arguments. Cross-country panel studies on e-commerce adoption have repeatedly shown the presence of time-invariant, unobserved heterogeneity at the country level, stemming from differences in cultural norms, institutional maturity, legal frameworks, and digital infrastructure [48,49,50]. These factors, while difficult to observe or quantify, are likely correlated with key explanatory variables and may bias RE estimates. The FE model accounts for this by controlling for all unobserved, time-invariant country-specific characteristics, thereby ensuring more reliable within-country estimates. This approach aligns with prior research that emphasizes the role of institutional and socio-economic embeddedness in shaping digital commerce trajectories [51]. Hence, the FE model is not only theoretically sound but also consistent with the existing literature on unobserved country heterogeneity in e-commerce contexts. The estimation summary for the Fixed-Effect model is presented in the table below, and in this model, the dependent variable is represented as the enterprises’ total turnover from e-commerce sales, and the independent variables are pay to advertise on the internet; use social networks; use the enterprise’s blog or microblogs; and use multimedia content sharing websites. This analysis uses a model with country (entity) fixed effects. There are 243 observations across 27 countries with an average of 9 years per country, shown in Figure 2.

R2 (Within) is about 0.13, indicating that the model explains around 13% of the variation in turnover within countries over time. R2 (Overall) is 0.0855, meaning about 8.6% of the total variance in turnover is explained by the model. The F-statistic value is 7.79 (p < 0.001), indicating that the model is statistically significant.

Parameter estimates for the fixed-effects model are presented in

Table 5. Fixed Effects—Parameter Estimates.

	Coefficient	Std.Err	T-Stat	p-Value	95% CI Lower	95% CI Upper
Const	7.2915	2.3578	3.0925	0.0023	2.6437	11.9393
Pay to advertise on the internet;	0.1583	0.067	2.3621	0.0191	0.0262	0.2905
Use social networks (e.g., Facebook, LinkedIn, Xing, Viadeo, and Yammer);	0.0741	0.025	2.9595	0.0034	0.0247	0.1235
Use the enterprise’s blog or microblogs (e.g., Twitter and Present.ly);	0.039	0.0891	0.4377	0.6621	−0.1367	0.2147
Use multimedia content sharing websites (e.g., YouTube, Flickr, Picasa, and SlideShare).	0.0151	0.0149	1.0128	0.3123	−0.0143	0.0445

Source: own calculations.

. The constant (intercept) is 7.29 and is the estimated average value of the enterprises’ total turnover from e-commerce sales when all independent variables are zero, after accounting for country fixed effects. The p-value of 0.0023 is less than 0.05, and the intercept is statistically significant at the 5% level. In the context of this fixed-effects model, the intercept represents the baseline level of e-commerce turnover for a country, after controlling for the effects of the independent variables and the country-specific fixed effects. The 95% confidence interval (2.6437, 11.9393) means a 95% confidence that the true intercept lies within this range.

For the variable “Pay to advertise on the internet”, the positive and significant coefficient (0.1583) and the p-value of 0.0191 (statistically significant at the 5% level) suggest that online advertising is associated with higher e-commerce sales. For each 1 percentage point increase in the share of enterprises that pay to advertise on the internet, the total turnover from e-commerce sales increases by about 0.16 units, holding other factors constant. The effect of the use of social networks (coefficient 0.0741; p = 0.0034) is statistically significant at the 1% level, indicating that social network engagement is positively linked to e-commerce performance. A one percentage point increase in the use of social networks by enterprises is associated with a 0.07 unit increase in e-commerce turnover.

For the variable “use the enterprise’s blog or microblogs (e.g., Twitter and Present.ly)”, the coefficient is positive (0.0390) but not statistically significant (p = 0.6621). This means there is no strong evidence that using blogs or microblogs by enterprises has a measurable effect on e-commerce turnover in this dataset. Similarly, the use of multimedia content sharing websites (β = 0.0151; p = 0.3123) does not show a statistically significant relationship with e-commerce turnover.

Activities like paying for online advertising and using social networks are both positively and significantly associated with higher e-commerce sales among enterprises, after accounting for country-specific effects. By contrast, using blogs/microblogs and multimedia sharing sites does not show a significant impact on e-commerce turnover in this analysis.

The scatter plot showing the relationship between the percentage of enterprises that pay to advertise online and their total turnover from e-commerce sales is presented in Figure 3. Each point represents a country–year observation and the red line is the fitted regression line with a 95% confidence interval.

In the previous figure, there is a clear positive trend: as more enterprises engage in online advertising, e-commerce turnover tends to increase. This visual pattern supports the regression results, showing a statistically significant positive effect of online advertising on e-commerce sales.

A scatter plot showing the relationship between the percentage of enterprises using social networks (Facebook, LinkedIn, etc.) and their total turnover from e-commerce sales is presented in Figure 4. The visualization reveals a positive relationship between social network usage and e-commerce turnover, which aligns with the regression findings. The positive trend line confirms that, on average, countries with higher rates of enterprise social network usage tend to have higher e-commerce turnover, supporting the hypothesis that social media engagement contributes to the success of online sales.

Several statistical tests were applied for the fixed-effects regression model: the Breusch–Pagan test for heteroskedasticity, the Woolridge test for serial correlation, and the Pesaran Test for cross-sectional dependence. The Breusch–Pagan test results (LM statistic = 8.45; LM p-value = 0.076; F-statistic = 2.15; F p-value = 0.0758) suggest that we failed to reject the null hypothesis of homoskedasticity (p-value = 0.076) so there is no strong evidence of heteroskedasticity in the model.

The results of the Woolridge Test for serial correlation (coefficient rho = 0.3346; test statistic = 4.9175; p-value = 0.00) shows a significant serial correlation (p < 0.001), meaning that the residuals are correlated across time periods within countries. The Pesaran CD test indicates strong cross-sectional dependence (p < 0.001), suggesting that shocks affecting one country also affect other countries. These violations of standard panel regression assumptions suggest the possibility of using other models and approaches, robust standard errors (clustered by country and/or time) and a dynamic panel model, to address serial correlation.

First, the fixed-effects model is re-estimated using robust standard errors, clustered by both country and year. This approach helps correct the standard errors for both types of correlation, making the inference more reliable. The results are presented in the Figure 5 below.

The value of R2 (within) of 0.0337 shows that the independent variables explain only 3.37% of the variation within each entity. R2 (between), with a value of 0.3744, shows that the model explains 37.34% of the variation in each entity (differences between countries). The value of R2 of 0.0146 shows that the model explains, in total, only 1.46% of the variation in the dependent variable. Regarding the statistical significance tests, F-statistic = 0.7579 (p-value = 0.9831) and F-statistic (robust.) = 0.0976 (p-value = 0.9831), with p-values well above the standard significance threshold (0.05), so we cannot reject the null hypothesis that all coefficients are 0. In practice, the model is not statistically significant even at the robust level.

The model explains very little of the variation in turnover from e-commerce sales. There is no statistical evidence that the included predictors have a relevant common effect. From this perspective, the data on parameter estimates included in

Table 6. Robust standard errors—Parameter Estimates.

	Coefficient	Std.Err	T-Stat	p-Value	Lower CI	Upper CI
Pay to advertise on the internet;	0.1125	0.0683	1.6474	0.1010	−0.0221	0.2471
Use social networks (e.g., Facebook, LinkedIn, Xing, Viadeo, and Yammer)	0.0060	0.0571	0.1060	0.9157	−0.1065	0.1186
Use the enterprise’s blog or microblogs (e.g., Twitter and Present.ly)	0.0772	0.1257	0.6140	0.5399	−0.1706	0.3250
Use multimedia content sharing websites (e.g., YouTube, Flickr, Picasa, and SlideShare)	−0.0077	0.0104	−0.7408	0.4597	−0.0282	0.0128

Source: own calculations.

are relevant.

None of the variables analyzed is, in this model, statistically significant at the 5% threshold, which suggests that, in the current form of the model, we cannot confidently state that these variables influence e-commerce turnover from e-commerce sales. The only variable with a slightly more pronounced effect is “Pay to advertise on the internet”, which presents a positive coefficient (0.1125) and has p = 0.1010. The other three variables analyzed do not have significant effects, which may suggest an inefficient use of these channels.

To further address the serial correlation, we estimated a dynamic panel model (Arellano–Bond style) by including a lagged dependent variable and using first-difference estimation. This helps account for the persistence in turnover over time and provides more robust results in the presence of serial correlation. The first differences in the dynamic panel model (Arellano–Bond style) are presented in the Figure 6 below.

The estimated model provides mixed results in terms of the quality of fit and the statistical significance of the regression. The negative value of R2 (within), for the internal variation in entities over time, shows that the model used does not efficiently explain the variation within the units of analysis (countries over time). The R2 (overall) value of 0.0882 shows that approximately 8.8% of the total variation in the dependent variable is explained by the independent variables (a relatively modest value, but better than in the first model analyzed). The F-test (F-statistic = 2.3557, p-value = 0.0422) is significant at the 5% level, which means that the model, as a whole, is statistically significant and that at least one independent variable has a significant effect on the dependent variable. The model is statistically significant at the global level, but has a modest explanatory power. Parameter estimates for the dynamic panel model (Arellano–Bond style) are presented in the

Table 7. Dynamic panel model (Arellano-Bond style)—Parameter Estimates.

	Parameter	Std.Err	T-Stat	p-Value	Lower CI	Upper CI
Pay to advertise on the internet	0.1565	0.0641	2.4413	0.0156	0.0300	0.2830
Use social networks (e.g., Facebook, LinkedIn, Xing, Viadeo, and Yammer)	−0.0057	0.0230	−0.2468	0.8053	−0.0510	0.0396
Use the enterprise’s blog or microblogs (e.g., Twitter and Present.ly)	0.0040	0.1372	0.0291	0.9768	−0.2667	0.2747
Use multimedia content sharing websites (e.g., YouTube, Flickr, Picasa, and SlideShare)	−0.0008	0.0086	−0.0964	0.9233	−0.0178	0.0161
Lag turnover	−0.1728	0.0923	−1.8730	0.0627	−0.3548	0.0092

Source: own calculations.

below.

The only statistically significant effect (at 5%) in the dynamic panel (first difference) model is for “Pay to advertise on the internet”. A one percentage point change in enterprises paying for internet advertising is associated with a 0.1565 percentage point change in e-commerce turnover. Other variables (use social networks, blogs/microblogs, multimedia content sharing) are not statistically significant. The lagged dependent variable (previous year’s turnover) shows a marginally significant negative effect, suggesting mean reversion. This implies that while these Web 2.0 tools can trigger performance bursts, due to temporary boosts in engagement, the long-term benefit remains in their cumulative strategic integration. These longer-term benefits may not be fully captured by short-term revenue dynamics.

5. Discussions

The results obtained in the case of the random-effects model and the dynamic panel model (Arellano–Bond) show that the effect of “paying to advertise on internet” on enterprises’ total turnover from e-commerce sales is positive and significant, which confirms the first research hypothesis and a large part of the results of previous studies [37,38,40]. Internet advertising has proven, in all its forms in the context of Web 2.0, to be a decisive tool for increasing the revenues of enterprises in the 27 countries of the European Union.

In addition, the research results (in the case of the random-effects model) confirm the second research hypothesis regarding the influence of social media use on enterprises’ total turnover from e-commerce sales (positive and statistically significant influence in the case of the random-effects model, positive but statistically insignificant influence in the case of the fixed-effects model). A discrepancy could be observed in the estimated effect of social media penetration across model specifications. Specifically, while the random-effects (RE) model suggests a significant positive impact (β = 0.0793, p < 0.01), this effect becomes statistically insignificant in both the fixed-effects (FE) model with clustered standard errors (β = 0.0060, p = 0.9157) and the dynamic panel model (β = −0.0057, p = 0.8053). This divergence likely reflects the presence of unobserved, time-invariant country-level heterogeneity—such as institutional quality, cultural attitudes towards technology, or policy frameworks—that is not accounted for in the RE model. When such heterogeneity is correlated with social media usage, RE estimates may be biased and overstate the significance of explanatory variables. By contrast, the FE and dynamic models better control for this bias by focusing on within-country variation over time. Therefore, although RE results suggest a statistically significant association, we interpret this with caution. Given the theoretical and empirical reasons discussed previously, we consider the FE-based findings to provide more robust evidence, especially in light of unobserved country-specific factors that may confound the relationship. The positive influence of social media on the development of e-commerce thus confirms some of the results of previous studies [6,45]. However, the research results only partially reflect the extent of the effects of social media on the development of e-commerce, revealed by other qualitative or quantitative approaches of previous studies in this regard.

A novel element of the research conducted, as one of the key findings, revealed by the results obtained is that “use the enterprise’s blog or microblogs (e.g., Twitter and Present.ly)” and “use multimedia content sharing websites (e.g., YouTube, Flickr, Picasa, and SlideShare)” do not influence enterprises’ total turnover from e-commerce sales to a valid and statistically significant extent. Thus, the third and fourth hypotheses are not confirmed by the results of the research conducted. In addition, the research results regarding “use the enterprise’s blog or microblogs (e.g., Twitter and Present.ly)” and “use multimedia content sharing websites (e.g., YouTube, Flickr, Picasa, and SlideShare)” do not support and do not agree with other, similar studies [39,40,43], which is explainable as European enterprises are not among the most important innovators in this field nor the owners of the most important multimedia content sharing websites (e.g., YouTube, Flickr, Picasa, and SlideShare). A JRC report written by the European Commission [52] shows that the EU contributes with only 10% in this segment of activity. Even if there is an ascending trending of innovation performance registered since 2017 [53], the majority of Web 2.0 instruments are developed by USA entities. Recent studies highlight the fact that the EU did not create large technological platforms comparable with the ones owned by the USA [54], due to its structural limitations related to entrepreneurship culture, governance, and financial support [55].

An inclusion of data from other geographical areas (North America, Asia) where e-commerce is even more developed compared to European countries would probably improve the chances of confirming these two research hypotheses.

The analysis of the obtained results and the comparison with other specialized studies show the theoretical and practical implications of the study presented in this article. The most important theoretical implications of our study concern the need to expand the role of perception on online advertising and the use of social media by European enterprises in the development of e-commerce. The amplification and re-conceptualization of online contacts with consumers in the context of Web 2.0 thus become essential from a theoretical perspective for the development of e-commerce. From a pragmatic perspective, of the practical implications of the research carried out, it has become evident the need to focus the relationship with both customers/consumers on certain components of the online environment such as social media in relation to others (microblogs). From the perspective of managerial practice, the results reconfirm and emphasize the need to restructure the commercial operational areas of contemporary organizations towards their virtual side, to increase the role of social media and social virtual networks in amplifying e-commerce and the organizations’ revenues from sales in the virtual space.

The study conducted and the explanatory and generalizable power of the results are affected by a series of limitations due to the geographical space from which the research data come and the fact that they mainly reflect the organizational approach and less so the individual one. Therefore, one of the future directions of this research is to extend it by considering data for other geographical spaces representative of the development of Web 2.0 and e-commerce.

6. Conclusions

This literature review showcases the role of the Web 2.0 tool for e-commerce development at the level of the European Union. Our study contributes to answering the questions identified in the gap in the literature, by using a regression analysis based on panel data gathered from Eurostat, which has the power to show relevant perspectives for the digital transformation of the EU, for the period of 2015–2023. By applying fixed-effects models and a dynamic Arellano–Bond model, this research provides robust empirical evidence of the real and measurable impact of Web 2.0 technologies on e-commerce revenue.

Theoretical and practical implications: This paper brings both theoretical and practical contributions, integrating the economic sustainability perspective for e-commerce development. From a theoretical point of view, this study shows the importance of interactivity between companies and consumers in the digital environment [44], using Web 2.0 tools. From a practical point of view, the results suggest the need to prioritize social media strategies and update commercial operations in order to respond to the needs of clients. The bibliometric review showed a lack of studies that applied the economic effectiveness of Web 2.0 tools on e-commerce.

The first two research hypotheses were confirmed through the existing literature and through the econometrical analysis, underlining the positive and statistically significant effect of internet advertising [42] and social media usage on the total turnover of e-commerce sales. On the other hand, blogs, microblogs, and multimedia content sharing platforms were not considered statistically significant, underlining the discrepancy between available technological resources and the actual integration in the digital economy. The main contribution of this research is related to providing quantitative outcomes, in a field dominated by qualitative outcomes or individual country, company, and industry approaches.

This study integrates Web 2.0 instruments into the econometric framework, unlike other studies which investigate only the determinants of e-commerce [18]. The combined techniques used are capturing both the dynamics of e-commerce and country heterogeneity, providing relevant perspectives for public policy, digital marketing, and companies in general. Among Web 2.0 tools, online paid advertising and the usage of social media are the ones contributing to the increase in e-commerce revenue. While the other variables, blogs, microblogs, and multimedia platforms are not considered relevant, the outcomes are confirmed by applying dynamic and robust models. The results highlight that online paid advertisement and social media usage variables have a significant, positive effect on e-commerce performance, confirming the first and second hypotheses. “Use the enterprise’s blog or microblogs” and “use of multimedia content sharing websites” do not influence enterprises’ total turnover from e-commerce sales to a valid and statistically significant extent. Thus, the third and fourth hypotheses are not confirmed by the results of the research conducted, possibly due to limited innovation and platform ownership in Europe. Web 2.0 instruments play an active role in the digital economy of the European Union, as it is considered a digital infrastructure with measurable economic effects.

Limitations and future examination: In the analysis of the impact of Web 2.0 tools on e-commerce at the level of the 27 member states, the contextual variables such as digital literacy, legislation, and cultural characteristics were omitted. The reason for this is related to our objective to proceed with the analysis at a macro level, using general variables available for all member states. In a future study, we can continue with the research by analyzing the contextual variables which could be gathered through questionnaires. Furthermore, another limitation is determined by the fact that this study does not differentiate between types of e-commerce businesses or product categories; our analysis placed greater emphasis on aggregated national-level data rather than on the particularities of business sectors and types of products. We acknowledge that multimedia content may be essential in sectors where product assembly or product usage is complex; therefore, future research could investigate such sector-specific effects. In addition, environmental impacts, such as reductions in carbon footprint, are beyond the scope of this study but remain relevant for future research. Future research can be extended by including the diversity of economic sectors and types of organizations, by including additional control variables and by exploring the transitions from Web 2.0 to Web 3.0.

The outcomes of the quantitative research validate the hypothesis related to the fact that Web 2.0 instruments, especially paid online advertising and the usage of social media, have a positive influence on the turnover generated by electronic commerce. This finding can be applied exclusively to the 27 EU member states and cannot be generalized. This study can serve as a starting point for international collaborative research, but it needs to be adapted to the regional realities. Therefore, this research can be extended to other regions like America and Asia, based only on an analysis of data collected directly from those regions. Future studies will enable comparative evaluations of the impact of Web 2.0 instruments in different regions.