353 Organizacija, V olume 58 Issue 4, November 2025 Research Papers 1 Received: 6th May 2025; Accepted: 21st August 2025 Business Analytics and Digitalization as Drivers of Startup Evaluation: The Experience of the Baltic States Valeriia SHCHERBAK 1 , Oleksandr DOROKHOV 2 , Kadri UKRAINSKI 2 , Deniss DJAKONS 3 , Olha KOVALYOVA 1 , Liudmyla DOROKHOVA 4 1 Sumy National Agrarian University, Department of Economic and Entrepreneurship, Sumy, Ukraine, valeriia.shcherbak@snau.edu.ua, olgakovalyovasumy@gmail.com 2 University of Tartu, Department of Public Economics and Policy, Tartu, Estonia, oleksandr.dorokhov@ut.ee, kadri.ukrainski@ut.ee 3 ISMA University of Applied Sciences, Riga, Latvia, deniss.djakons@isma.lv 4 University of Tartu, Department of Marketing, Tartu, Estonia, liudmyla.dorokhova@ut.ee Purpose: This study is motivated by the importance of startups in economic growth and the need for methods to evaluate their success, considering risk and uncertainty. The objective is to analyze factors that influence startups, using factor and cluster analysis. The hypothesis that advanced business analytics in startup evaluation can en- hance the quality of investment decision-making was tested. Methods: The combination of quantitative and qualitative techniques was used. Statistics about 20 startups from Latvia, Lithuania, and Estonia over five years were processed to identify success drivers and to group startups by similarity. Machine learning and social media sentiment analysis were applied to assess non-financial indicators. Results: The results showed that indicators such as projected profitability, social media activity, and innovativeness are significant for startup ranking. The share of traditional methods in the Baltic states was 55%, while modern tools were 45%, highlighting the role of digitalization in risk assessment. Startups with high clustering coefficients and positive mention sentiment demonstrated superior performance. Conclusions: The study demonstrated that integrating business analytics and digitalization enhances startup eval- uation. The model combines financial metrics with network and sentiment analysis, offering a comprehensive frame - work for investors. It confirms that data-driven methods improve decision-making, reducing investment risks. Keywords: Startup evaluation, Business analytics, Digitalization, Baltic States, Economic potential, Social engage- ment DOI: 10.2478/orga-2025-0022 1 Introduction The current conditions of high uncertainty and dyna- mism in the business environment require effective deci- sion-making approaches from investors and organizations, especially in the field of startup financing. Startups play a key role in the innovative development of the economy, creating new jobs, developing technolo- 354 Organizacija, V olume 58 Issue 4, November 2025 Research Papers gies, and contributing to market competitiveness. Start- ups play a key role in the innovative development of the economy, creating new jobs, developing technologies, and contributing to market competitiveness (Startup Genome, 2025). However, investing in startups is associated with high risks due to their limited operating history, uncertainty of market success, and insufficient information about future development. Startup financing decisions require a com- prehensive approach that considers both risks and oppor- tunities. Traditional analysis methods are often insufficient for assessing startup potential, which increases interest in us- ing data-driven analytical tools. Business analytics meth- ods, including descriptive, diagnostic, predictive, and pre- scriptive analytics, offer new opportunities for evaluating startups and making more informed decisions. This article is dedicated to studying approaches to bal- ancing risk and opportunity in startup investing. Particu- lar attention is paid to the application of modern business analytics methods, including machine learning, network science, and social media analysis. The aim of the research is to develop and evaluate an- alytical tools that will help investors make more accurate and objective startup financing decisions, contributing to their success and growth. This work contributes to the development of theo- retical and practical knowledge about the application of business analytics in investment activities, offering new perspectives for supporting innovation and sustainable de- velopment. However, previous studies have mostly focused on in- dividual aspects of startup evaluation, such as access to finance (Fisch, 2018), innovation performance (Kim et al., 2024), or the application of business analytics in SMEs (Anuradha and Sailaxmi, 2024). Only limited research has addressed the integration of advanced analytical tools — machine learning, network analysis, and social media diagnostics — for comprehen- sive startup evaluation. The research gap addressed in this study lies in the absence of a unified framework that combines traditional financial metrics with digital indicators (e.g., social media activity, network centrality) for startup evaluation. Moreover, regional studies on the Baltic States remain scarce, despite the region’s growing importance as a hub for innovative startups (Startup Genome, 2025; LSM, 2025). Our work bridges this gap by developing and testing a hybrid multifactor model that integrates economic, tech- nological, and social indicators, thus contributing both to academic literature and to practical investment deci- sion-making in the context of the Baltic startup ecosystem. 2 Literature overview 2.1 Financing and survival of startups and SMEs An analysis of available financing sources for startups and small and medium-sized enterprises (SMEs), as well as a study of the factors determining their survival and success, has shown that the sustainable development and financial stability of these organizations play a key role in economic growth, innovation, and job creation. Traditional sources of financing include bank loans, which remain the primary financial instrument for many SMEs. However, research by Calabrese and Osmetti (2013) emphasizes the high risks of default, especially in the case of rare but significant events. The use of a gen- eralized extreme value regression model allows for a de- tailed analysis of the probabilities of such risks. The study by Coleman et al. (2016) examines US startups’ decisions regarding debt financing. This research helps identify fi- nancing structures and their impact on startup financial sta- bility, providing empirical data on the factors influencing the successful use of debt. Alternative financing sources, such as crowdfunding and venture capital, are becoming increasingly popular (Agrawal et al., 2014). Tomczak and Brem (2013) con- ceptualize the crowdfunding investment model, focusing on its role in diversifying startup financing sources. Teker et al. (2016) analyze venture capital markets, providing a cross-country analysis of venture capital availability for startups. The importance of non-financial information for credit risk assessment is highlighted in the work of Wahl- strøm et al. (2024). The integration of such data improves financing decision-making processes, especially in the context of SMEs (Gazzola et al., 2022). Alternative financing for SMEs in the Baltic states, ac- cording to Rupeika-Apoga (2014), represents a significant source of financial resources. The study by Fisch (2018) focuses on the differences in access to alternative financing sources across different regions. Factors influencing the longevity and success of startups are detailed in research by Keogh and Johnson (2021). Econometric analysis allows for the identification of such aspects as financing structure, access to capital markets, level of competition, and the adaptability of busi- ness models (Foreman-Peck et al., 2006). Thus, the diversity of financing sources and an un- derstanding of the survival factors of startups and SMEs require a comprehensive approach. This will allow for ef- fective assessment of their financial stability and the de- velopment of strategies aimed at long-term success and growth. 355 Organizacija, V olume 58 Issue 4, November 2025 Research Papers 2.2 Innovation and SME growth Innovations have a significant impact on the growth and development of small and medium-sized enterprises (SMEs), with an emphasis on financial constraints, re- gional characteristics, and cooperative research and de- velopment (R&D). Financial constraints are a key barrier to SME innovation activity (Chatterji et al., 2018). Acebo et al. (2020) note that innovation subsidies can partially compensate for these constraints, stimulating investment in R&D. However, the effect of subsidies varies depending on the level of financial accessibility: for firms with limited access to capital, such subsidies have a more significant impact (Ciampi and Gordini, 2012). This underscores the need for government support for innovation, especially in the context of tight financial constraints. The regional context plays an important role in the development of innovation activity in medium-sized busi- nesses. Research by Berlemann and Jahn (2015) empha- sizes that medium-sized firms in regions with high levels of infrastructure and access to scientific resources demon- strate higher innovation efficiency. This is explained by the presence of local ecosystems that facilitate knowledge sharing and technological breakthroughs. Thus, territorial characteristics should be taken into account when develop- ing SME support strategies. Cooperative R&D is a powerful tool for increasing SME innovation activity. Research by Kim et al. (2024) demonstrates that collaboration between firms, universi- ties, and research institutions contributes to accelerating the development of new technologies and products. The example of South Korean SMEs in the manufacturing sec- tor shows that participation in cooperative R&D not only increases the competitiveness of companies but also re- duces the risks associated with innovation activities. Entrepreneurial activity and innovation are key factors for economic growth. Wong et al. (2005), in their research based on Global Entrepreneurship Monitor (GEM) data, emphasize that a high level of innovation in the entrepre- neurial environment leads to accelerated economic devel- opment. At the same time, SMEs play an important role, contributing to job creation and technology development. 2.3 Business analytics and digitalization for SMEs Business analytics and digitalization play a crucial role in the transformation of small and medium-sized enterpris- es (SMEs), contributing to increased competitiveness, ef- ficiency, and adaptability (Melegati et al., 2019). Business analytics tools, such as Growth hacking, provide a target- ed approach to business process optimization (O’Neill and Brabazon, 2019). Research by Anuradha and Sailaxmi (2024) demon- strates how the use of such tools helps SMEs achieve growth by analyzing consumer behavior, increasing the profitability of marketing campaigns, and improving data management. Al-Debei (2023) emphasizes the importance of clearly distinguishing between the concepts of business analytics and data science. Recent research also highlights the global role of AI and digital technologies in shaping IT startup ecosystems (Hemanth and Lakshminarayana, 2025) and in promoting sustainable innovation in green startups (Fichter et al., 2025). Business analytics focuses on the practical application of data to improve decisions, while data science includes the development of complex models and algorithms. This distinction allows SMEs to effectively choose appropriate methods for their goals. Baijens et al. (2021) propose a theoretical model for data analytics management based on the VSM (Viable System Model). This model helps SMEs effectively structure data processing, ensuring flexibility and resilience to change. Research by Ioakeimidou et al. (2024) presents a new measurement scale for assessing data analytics maturity. This tool allows SMEs to determine their current level of analytics development and formulate strategic plans to achieve a higher level of digital maturity. AI-driven tools for startup evaluation are increasingly discussed in the context of data analytics and investment decision-making (Lutfiani et al., 2025). Kato et al. (2023) explore how the selection of relevant information affects the effectiveness of analytics. Using redundant informa- tion can reduce the quality of decisions, so it is important to identify key data for evaluating sales and testing con- cepts. This trend is consistent with global findings on the evolution of IT startup ecosystems under the influence of AI (Hemanth and Lakshminarayana, 2025). Research by Qin et al. (2022) analyzes the demand for business analyt- ics skills in various industries. This allows SMEs to adapt their analytical strategies, focusing on labor market needs and developing employee competencies in the most in-de- mand areas. Quansah (2024) emphasizes that the imple- mentation of digital technologies is often associated with barriers, especially in low-income countries. Nevertheless, digitalization is becoming a necessary element for improving operations, expanding markets, and increasing competitiveness. Yaakobi et al. (2019) demon- strate how machine learning methods can be used to evalu- ate and optimize organizational projects. Machine learning methods, including random forest and gradient boosting algorithms, allow for the analysis of a wide range of fac- tors affecting performance (Blanquet et al., 2025). This is especially relevant for SMEs, which need to improve the efficiency of their operations and reduce management costs. 356 Organizacija, V olume 58 Issue 4, November 2025 Research Papers 2.4 Regional aspect and internationalization of SMEs The development of small and medium-sized enter- prises (SMEs) is determined by both regional factors and their ability to access international markets. Regional net- works, capital structure, financial institutions, and interna- tionalization all influence SME growth and sustainability (Kaya and Persson, 2019). Research by McAdam et al. (2015) emphasizes the im- portance of horizontal regional networks in the agri-food sector. Such networks stimulate knowledge sharing, col- laboration, and innovation among SMEs. This is particu- larly important in sectors where business success depends on joint actions, such as market access, production inno- vation, and supply chain resilience. Regional financial in- stitutions play a crucial role in providing capital to SMEs. Palacín-Sánchez and Di Pietro (2015) demonstrate that capital availability through regional banks and credit in- stitutions influences SME capital structure. In regions with a developed financial sector, companies are more likely to use long-term investment strategies, while in less devel- oped regions, short-term loans prevail. SME development depends on local policies, including the provision of sub- sidies, tax breaks, and support programs. Regional governments play a key role in creating con- ditions for sustainable growth and enhancing SME com- petitiveness. The work of Wright et al. (2007) emphasizes that internationalization allows SMEs to access new mar- kets, diversify revenues, and increase their competitive- ness. International entrepreneurship promotes innovation, technology transfer, and the development of business re- lationships. The main barriers to SME entry into internation- al markets include limited financial resources, a lack of knowledge about target markets, and weak infrastructure. These barriers are particularly significant for companies operating in regions with low levels of economic activity. Internationalization also depends on the ability of SMEs to adapt to different political and cultural contexts. This re- quires the development of flexible strategies and the use of local partners to minimize risks. Research by Sutherland et al. (2019) indicates that employers and regional partner- ships play a key role in supporting SME internationaliza- tion through training, practical assistance, and “try before you buy” programs. This approach reduces the risks as- sociated with entering new markets and promotes gradual integration into the global economy. 2.5 Incubators, networks, and university- business interactions The support infrastructure for small and medium-sized enterprises (SMEs), including business incubators, region- al networks, and university-business interaction, plays a crucial role in the development of innovative entrepre- neurship, knowledge transfer, and personnel training. According to Aernoudt (2004), business incubators provide startups with infrastructure, mentorship, and ac- cess to funding. They help new businesses overcome bar- riers in the initial stages, creating favorable conditions for their growth and sustainability. Incubators act as catalysts for innovation, promoting accelerated business develop- ment through access to resources and supporting ecosys- tems. Key success factors for incubators include the availa- bility of quality mentorship, active involvement of partners from business and academia, and ensuring the accessibil- ity of financial instruments. Incubators also contribute to the development of entrepreneurial skills, which increases SME competitiveness in the market. Research by McAd- am et al. (2015) emphasizes the importance of horizontal regional networks for stimulating innovation in the agri- food sector. Such networks create a platform for the ex- change of experience and knowledge among participants, contributing to the development of the local economy and enhancing SME competitiveness. Successful regional networks are characterized by a high degree of involvement of all stakeholders, includ- ing business, universities, and government organizations. They play a key role in addressing specific regional chal- lenges, such as access to resources and the adaptation of innovative solutions. Dada et al. (2015) explore the fran- chising of university-business interaction as an effective tool for knowledge and technology transfer. Universities can contribute to SME development through training programs, research projects, and intern- ships. This interaction is particularly important for training qualified personnel who meet business needs. The impact of human capital on SME development is emphasized in the work of Sutherland et al. (2019). In- ternational student mobility provides a unique experience that can be used for the development of local enterprises. Students with international experience bring new knowl- edge and approaches, which contribute to innovation and the strengthening of ties between universities and busi- nesses. 2.6 Entrepreneurship in times of crisis and special groups of entrepreneurs In times of crisis, entrepreneurship plays an important role as a mechanism for adaptation and economic recov- ery. Support for entrepreneurship among specific groups, such as refugees, who face unique challenges and opportu- nities, becomes particularly important. Research by Bizri (2017) focuses on the role of so- cial capital in refugee entrepreneurship. Social networks, 357 Organizacija, V olume 58 Issue 4, November 2025 Research Papers ties with diasporas, and community support are important factors helping refugees overcome barriers such as a lack of financial resources, language difficulties, and a lack of knowledge about local markets. Social capital not only stimulates business start-ups but also creates conditions for their sustainability and growth. The work of Kolodiziev et al. (2024) analyzes the contribution of refugee-founded startups to the economies of host countries. Such startups contribute to job creation, expansion of local markets, and stimulate the development of new business models. The authors emphasize that the successful integration of refugee entrepreneurs is possible with access to funding, training programs, and support from local authorities. Refugees face a number of unique barriers: lack of ac- cess to finance, linguistic and cultural differences, as well as restrictions in market access. These problems require targeted policies and support programs, including integra- tion into the entrepreneurial ecosystem of host countries. Economic and social crises often become catalysts for the emergence of new business ideas. In such conditions, en- trepreneurs are forced to adapt, develop innovative prod- ucts and services that meet changing market needs. During crises, SMEs play a key role in maintaining economic activity and creating jobs. Such enterprises pos- sess the flexibility to adapt quickly to changes and are able to effectively use local resources to meet demand. To sup- port entrepreneurship in times of crisis, it is necessary to implement financial assistance programs, tax breaks, and educational initiatives. Such measures stimulate the crea- tion of new enterprises and strengthen their sustainability in the long term. 2.7 Forecasting and Evaluation of SME Performance Forecasting the financial condition and assessing the performance (e.g., profitability, growth, operational effi - ciency) of small and medium-sized enterprises (SMEs) are key elements of their sustainable development. Research by Ciampi and Gordini (2012) demonstrates how artificial neural networks can be applied to forecast the probability of default for small businesses. These methods allow for the analysis of complex non-linear relationships between financial indicators and risk factors, making them a more accurate tool compared to traditional statistical models. The example of Italian small businesses shows that such approaches improve the predictive accuracy and help identify vulnerable enterprises at early stages. Jabeur and Fahmi (2017) conduct a comparative study of various fi- nancial distress forecasting models for French firms. The authors identify logistic regression as one of the most ef- ficient methods due to its simplicity and interpretability. However, it is emphasized that modern tools, such as neural networks and decision trees, demonstrate better per- formance on complex data. The article by Lu (2019) ana- lyzes the use of Bayesian estimation to improve the pre- dictive performance of logistic regression. This approach allows for considering the variability of predictors, which is especially important for forecasting SME financial sta- bility. Bayesian methods make models more adaptable to changes in data, which increases their practical applica- bility. Yaakobi et al. (2019) consider the application of machine learning methods for evaluating organizational performance. These methods, including random forest and gradient boosting algorithms, allow for the analysis of a wide range of factors affecting business outcomes. Machine learning can also be used to identify hidden patterns in data, which helps improve operational process- es and strategic planning. The assessment of KPIs, such as profitability, liquidity, and operational efficiency, is an integral part of SME management. Modern analytical tools integrate machine learning and statistical models to pro- vide more accurate and timely data for management deci- sion-making. 2.8 Research hypothesis and proof tasks The literature review in Sections 2.1–2.7 reveals two critical gaps in startup evaluation methodologies, men- tioned below. Overreliance on traditional financial metrics (Cal- abrese and Osmetti, 2013; Sivicka, 2018) often fails to capture non-financial drivers of success (e.g., social media engagement, network centrality). Limited integration of advanced analytics (e.g., ma- chine learning, sentiment analysis) into holistic frame- works, despite their proven accuracy in risk assessment (Ciampi and Gordini, 2012; Yaakobi et al., 2019). Recent studies (Hemanth and Lakshminarayana, 2025; Lutfiani et al., 2025) underscore the promise of hybrid models, yet they lack empirical validation in alternative contexts—such as the Baltic states. This study bridges the gap by proposing a unified approach that combines financial, technological, and social indicators, addressing the need for data-driven decision-making noted by Fisch (2018) and Rupeika-Apoga (2014). Research Hypothesis H1: “A comprehensive approach to risk and opportunity analysis using business analytics methods, such as ma- chine learning, network analysis, and social media diag- nostics, contributes to improving the quality of investment decisions in startups, increasing their chances of sustaina- ble development and market success.” Research Objectives: • Analysis of current approaches to startup risk 358 Organizacija, V olume 58 Issue 4, November 2025 Research Papers assessment. To achieve this objective, it will be necessary to conduct a review of traditional and modern methods of risk and opportunity analysis in startup investing; identify the limitations of tra- ditional approaches and the need for the imple- mentation of analytical tools. • Development of an analytical model for startup evaluation. To achieve this objective, it will be necessary to create a model that integrates ma- chine learning, network analysis, and social me- dia analysis methods to assess the prospects of startups; to test the effectiveness of the model on real data. • Evaluation of the impact of implementing an- alytical methods on the quality of investment decisions. To achieve this objective, it will be necessary to conduct a comparative analysis of in- vestment decisions made using the proposed mod- el and decisions based on traditional approaches, to assess the impact of the model on startup suc- cess indicators such as survival, profitability, and growth. • Identification of factors influencing startup suc- cess. To achieve this objective, it will be necessary to use the proposed model to identify key factors determining startup sustainability and market suc- cess, and to compare the results with previously identified factors in the literature. • Development of recommendations for investors. To achieve this objective, it will be necessary to formulate recommendations on the use of analyt- ical tools to minimize risks and maximize oppor- tunities in startup investing; to propose practical measures to improve the investment process. Expected Results: It is assumed that the use of modern analytical tools will improve the accuracy of assessing startup risks and opportunities, reduce the likelihood of erroneous invest- ment decisions, and contribute to the development of a more sustainable investment ecosystem that supports in- novation and economic growth. These objectives are aimed at proving the hypothesis about the importance of integrating analytical methods into the startup financing decision-making process, which has practical and theoretical significance for the develop- ment of investment activities. Data collection and the research itself were conducted from 2022 to 2024 in the Baltic states: Latvia, Lithuania, and Estonia. 3 Materials and Methods 3.1 Analysis of current approaches to startup risk assessment In the Baltic states, startups play a key role in economic development, acting as engines of innovation and job crea- tion. However, their financing is associated with high risks due to limited operating history, high market volatility, and a lack of information about future prospects. The conserv- ative approach to risk management in Latvia may be relat- ed to limited digitalization and a habit of using time-tested methods (LSM, 2025; Stats and Market Insights, 2025а; 2025b). An analysis of the advantages and disadvantages of traditional methods is presented in Table 1. Table 1 reveals that financial analysis is based on the analysis of balance sheet indicators such as profitability, li- quidity, and debt ratio. Its advantages lie in the ease of ap- plication and the possibility of using historical data; its dis- advantages lie in the limited applicability to startups due to the lack of extensive financial history. Expert assessments allow for risk evaluation based on expert opinions. Their advantages lie in the intuitive nature of the approach; the disadvantages lie in subjectivity and dependence on ex- pert qualifications. SWOT analysis is used to identify the strengths and weaknesses of startups, and opportunities and threats. Its limitations lie in the subjectivity of quanti- tative assessment. An analysis of modern risk assessment methods is presented in Table 2. Table 1: Advantages and disadvantages of startup risk assessment methods Method Advantages Disadvantages Financial analysis Based on objective data (financial statements), it allows for assessing financial stability and profitability. Limited availability of financial information for startups does not take into account non-financial factors. Expert assessments Takes into account the experience and knowledge of experts in the industry, allowing you to assess qualitative factors. Subjectivity, difficulty of scaling, and depen- dence on the qualifications of experts. SWOT analysis Allows a comprehensive assessment of strengths and weaknesses, opportunities and threats, and takes into account the strategic context. Subjectivity of assessments, difficulty of quantitative assessment of factors. Source: (Sivicka, 2018) 359 Organizacija, V olume 58 Issue 4, November 2025 Research Papers Table 2: Comparison of modern startup risk assessment methods Method Application area Advantages Disadvantages Machine Learning (ML) Forecasting, classification, clustering, big data analysis, and identifying patterns. High forecast accuracy with sufficient data, ability to self-learn and adapt to new data, and automation of pro- cessing large volumes of information. Requires large volumes of high-qual- ity data for training, difficulty interpreting results (“black box”), susceptibility to overfitting, and requires qualified specialists. Social Media Analytics Reputation assessment, public opinion analysis, identifying trends, and monitoring compet- itors. Real-time public opinion, the ability to identify potential crises at an early stage, and obtaining information about customer preferences. Limited data (availability, reliability), difficulty analyzing unstructured data (texts, images), susceptibility to manipulation. Network Analysis Assessing connections and influ- ence within a startup and in the external environment (investors, partners, clients), identifying key players and opinion leaders. Visualization and analysis of complex relationships, identification of hidden patterns, and potential risks associated with dependence on individuals or groups. The complexity of collecting and processing data on connections and the difficulty of interpreting complex network structures require special- ized software. Bayesian Approach Assessing uncertainty and the probability of various events, taking into account a priori knowledge and updating it with new information. Flexibility, ability to take into account subjective expert assessments, adaptability to changes, and ability to update forecasts as new data arrives. High complexity of calculations, need to determine a priori probabilities, results depend on the correctness of a priori estimates. Source: Authors’ aggregation based on (Brecht et al., 2021; Ciampi and Gordini, 2012; Yaakobi et al., 2019; Anuradha and Sailaxmi, 2024; McAdam et al., 2015; Lu, 2019) Table 3: Methods for startup risk assessment in Baltic states Country Traditional methods (%) Modern/ analytical methods (%) Specific methods used Comments Latvia 60 40 SWOT analysis, financial ratio analysis, ex- pert judgment; analytical methods include regression models and decision trees. Dominance of traditional methods reflects a conservative approach to risk assessment. Estonia 55 45 Scenario analysis, cash flow forecasting; advanced methods include machine learning algorithms and Monte Carlo simulations. Active use of analytical tools indicates a focus on compre- hensive and data-driven risk analysis. Lithuania 50 50 Break-even analysis, sensitivity analysis; modern tools include big data analytics and predictive modeling techniques. Balanced use of both approach- es suggests a preference for combining simplicity with precision. Baltic average 55 45 Weighted average of the methods across all countries. On average, the Baltic states exhibit a slight preference for traditional methods, though the gap with modern techniques is narrowing. Source: (EU-Startups, 2023; Liu et al., 2022) Table 2 highlights that machine learning is mainly used for forecasting the probability of default, analyzing market data, and customer behavior. An example is the application of classification methods (decision trees, neural networks). Social media analysis is used to study startup reputa- tion, user reviews, and market interest. Network analysis is used to identify partnerships and the startup’s market influence. The Bayesian approach is used to account for uncertainty in risk assessment. 360 Organizacija, V olume 58 Issue 4, November 2025 Research Papers Table 3 provides a structured overview of startup risk assessment approaches in the Baltic states (Latvia, Estonia, and Lithuania), including the distribution of traditional and modern methods, specific tools used, and commentary. In Latvia, 60% of traditional methods and 40% of mod- ern analytical approaches are applied. Simple tools such as SWOT analysis, financial ratios, and expert assessments prevail. The conservative approach to risk management may be related to limited digitalization and a habit of using time-tested methods. Latvia, with its dominance of tradi- tional methods, may face limitations in managing complex and dynamic risks, which puts it in a vulnerable position in global competition. In Estonia, 55% of traditional methods and 45% of modern methods are used. Scenario analysis and cash flow forecasting are widely used, as are advanced tools such as machine learning algorithms and Monte Carlo simulations. The use of analytical tools reflects the country’s high digital maturity and focus on innovation. Estonia stands out for its focus on comprehensive data analysis. Estonia demonstrates clear leadership in the ap- plication of modern approaches, which contributes to the formation of a more sustainable startup ecosystem. Lithuania shows an even distribution: 50% traditional methods and 50% modern assessment methods. Break-even and sensitivity analysis are mentioned, as well as advanced tools such as big data analytics and pre- dictive modeling. The balance between approaches indi- cates an attempt to combine the accessibility of traditional methods with the accuracy of modern technologies. Lithuania, thanks to its balanced approach, has the potential to integrate the best practices of both systems, which strengthens its position as a developing innovation center. The average for the Baltics is 55% traditional methods versus 45% modern methods. This reflects a slightly predominant role of traditional approaches, but the gap is narrowing due to the introduc- tion of modern analytical methods. 3.2 Developing an analytical model for evaluating startups The model for assessing the prospects of startups using the taxonomy method, machine learning, network analy- sis, and social media analysis is presented in Table 4. Table 4: Methodology for assessing startups 361 Organizacija, V olume 58 Issue 4, November 2025 Research Papers Source: Author’s methodology, based on (Foster, 2004; Murphy, 2012; Langfelder & Horvath, 2008; Anstead and O’Loughlin, 2014) Table 4: Methodology for assessing startups (continues) Figure 1: Dynamics of the number of startups in the Baltic countries (2018–23) Source: (Startup Lithuania, n.d.; Dealroom Database - Everyone Is Here - Startup Lithuania, 2022; EU-Startups, 2023; Startup Estonia, 2023) 362 Organizacija, V olume 58 Issue 4, November 2025 Research Papers The initial data and their symbols are given in the Ap- pendix. 4 Results 4.1 Startup ecosystem growth trends in the Baltic States (2018–2023) Figure 1 presents the quantitative evolution of startups across Lithuania, Latvia, and Estonia, revealing distinct sectoral and regional patterns that reflect the region’s in- novation landscape. Based on the data in Figure 1, the following conclu- sions can be drawn. All three countries demonstrate steady growth in the number of startups across all sectors during the observation period. This indicates a favorable environment for innovation in the Baltic states, which is associated with active govern- ment support and an increase in investment inflows. Estonia demonstrates the largest growth in startups in the IT sector (from 120 to 360) and other industries (from 360 to 720). This is due to a developed digital infrastruc- ture, access to international markets, and the country’s fo- cus on IT solutions. Latvia and Lithuania show significant growth in the fintech sector, especially in Lithuania (from 342 to 462). This may be due to attractive conditions for financial tech- nologies, including regulatory sandboxes and access to the European market. Green energy is developing in all countries, but Esto- nia is leading (from 240 to 280). This is due to the growing interest in sustainable technologies and the Baltic states’ desire to reduce their carbon footprint. In some sectors, for example, in green energy in Latvia and Lithuania, there is a slowdown in growth or even a decline (for example, in Latvia from 200 to 156). This may be due to limited funding or high barriers to market entry. The dynamics of startups in the Baltic states reflect their focus on technological development, with an empha- sis on IT, fintech, and green technologies. Estonia continues to lead due to its developed digital ecosystem, while Latvia and Lithuania demonstrate poten- tial in specific niches. This data underscores the importance of further supporting the innovation ecosystem through in- vestment, education, and international cooperation. Table 5: Results of factor analysis of the influence of individual variables on the ranking of startups in the Baltic States (2023) Variable Factor Loadings (Unrotated) (Data_nor) Extraction: Principal components (Marked loadings are > 700000) Factor 1 Factor 2 Х1 0,986012 0,055934 Х2 0,073367 0,990375 Х3 0,990215 0,020135 Х4 0,961665 0,050260 Х5 -0,056516 0,890580 Х6 0,096334 0,991384 Х7 0,967420 0,189275 Х8 0,036298 0,995327 Х9 0,020271 0,993515 Х10 0,035642 0,996444 Х11 0,762896 0,419920 Х12 -0,062001 0,691322 Х13 0,792666 0,192932 Expl.Var 6,758310 4,750997 Prp.Totl 0,550639 0,334692 where Х1 – Projected profitability, million €; Х2 – Activity in social networks, thousand subscribers; Х3 – Availability of investors, number; Х4 – Innovativeness of technologies, scores 1-10; Х5 – Basic level of social responsibility, score 1-10; Х6 – Number of links, node degree; Х7 – Cluster coefficient; Х8 – Betweenness centrality; Х9 – Number of mentions; Х10 – Sentiment of mentions; Х11 – Total Raised, M$; Х12 – Total Raised, M$; Х13 – Number of employees, thousand people. Source: Author’s calculations 363 Organizacija, V olume 58 Issue 4, November 2025 Research Papers 4.2 Evaluation of the importance of factors for ranking startups To analyze the factors influencing the success and de- velopment of startups in the Baltic region, information was collected on a number of companies. Table 5 contains data on 20 startups from Lithuania, Latvia, and Estonia, covering a wide range of indicators, from projected profitability and social media activity to the amount of investment raised and team size. This data serves as the basis for further research and the identifica- tion of key determinants of startup success. Based on the presented results of the factor analysis (Table 5), two factors can be identified that determine the ranking of Baltic startups. Factor loadings that are high- lighted in red influence the process; those that remain black do not. Factor 1, “Financial and Resource Potential and Inno- vativeness,” includes the following indicators with high loadings: X1 (0.986012): Projected profitability, million €; X3 (0.990215): Availability of investors, number; X4 (0.961665): Innovativeness of technologies, scores 1-10; X7 (0.967420): Cluster coefficient; X13 (0.792666): Num- ber of employees, thousands of people. This factor combines characteristics related to the fi- nancial condition, investment availability, level of innova- tion, and organizational structure of startups. Factor 2, “Social and Network-Reputational Activi- ty,” includes the following indicators with high loadings: X2 (0.990375): Activity in social networks, thousands of subscribers; X5 (0.890580): Basic level of social re- sponsibility, score 1-10; X6 (0.991384): Number of links, node degree; X8 (0.995327): Betweenness centrality; X9 (0.993515): Number of mentions; X10 (0.996444): Senti- ment of mentions. This factor describes the social activity of startups, their participation in network structures, and the level of media mentions. Regression equations for each factor are constructed using the significant variables: Factor 1: F1=1/6,758(0,986 ⋅X1+0,990 ⋅X3+0,962 ⋅X4+0,967 ⋅X- 7+0,793 ⋅X13) (7) Factor 2: F2=1/4,751(0,990 ⋅X2+0,891 ⋅X5+0,991 ⋅X6+0,995 ⋅X- 8+0,994 ⋅X9+0,996 ⋅X10) (8) Factor 1 explains 55.06% of the variance. Factor 2 explains 33.47% of the variance. In total, the two factors together explain 88.53% of the total variance, which indi- cates the high informativeness of the analysis. 4.3 Grouping of Baltic startups by growth potential and attracted investments The analysis of the structure of Baltic startup clusters for 2023 was made taking into account only the significant indicators identified by regression analysis (Figure 2). Figure 2: Results of the cluster analysis of Baltic startups (STATISTICA 13) Source: Author’s calculations 364 Organizacija, V olume 58 Issue 4, November 2025 Research Papers Table 6 presents the results of the cluster analysis per- formed in STATISTICA 13, demonstrating the composi- tion of the first cluster and the distance of each startup to its center. Cluster 1 includes five startups: Green Genius, Origin, Roibox, Naco, and Cenos. The analysis of distances to the cluster center (Table 6) shows that Origin (0.1805555) and Roibox (0.2055702) are closest to the center, indicating their high similarity to the typical characteristics of the cluster. Naco (0.2611897) and Green Genius (0.2834758) demonstrate slightly greater distances, and Cenos (0.3678699) is farthest away, indicating its lowest typical- ity for this group. The startups included in Cluster 1 are characterized by average or slightly above average values for most indica- tors related to profitability, social media activity, investor attraction (number), innovativeness, social responsibility, network indicators, and media influence. At the same time, they demonstrate below-average indicators for the amount of investment raised (Total Raised) and the number of em- ployees. Overall, Cluster 1 represents locally oriented startups demonstrating moderate development indicators and limit- ed resources, which distinguishes them from the larger and faster-growing companies represented in other clusters. Table 7 presents the composition of the second clus- ter obtained as a result of cluster analysis in STATISTICA 13, and the distances of each startup to the center of this cluster. Cluster 2 includes four startups: Vinted, Aerones, Ovoko, and Sonarworks. The analysis of distances to the cluster center (Table 7) shows that Aerones (0.267903) and Sonarworks (0.315484) are relatively close to the center, demonstrating greater similarity within the group. Ovoko (0.397701) and especially Vinted (0.591298) are located further away, indicating their greater variability relative to the typical characteristics of the cluster. The startups included in Cluster 2 are characterized, on average, by below-average indicators for the sample across most criteria related to profitability, social media activity, investor attraction (number), network indicators, and media influence. Table 6: Composition of the 1 cluster (STATISTICA 13 cluster analysis listing) Table 7: Composition of the 2 cluster (STATISTICA 13 cluster analysis listing) Table 8: Composition of the 3 cluster (STATISTICA 13 cluster analysis listing) Members of Cluster Number 1 (Data_nor) and Distances from Respective Cluster Center Cluster contains 5 cases Case No. Distance Case No. Distance Green Genius 0,2834758 Naco 0,2611897 Origin 0,1805555 Cenos 0,3678699 Roibox 0,2055702 Source: Author’s calculations Members of Cluster Number 2 (Data_nor)and Distances from Respective Cluster Center Cluster contains 4 cases Case No. Distance Case No. Distance Vinted 0,591298 Ovoko 0,397701 Aerones 0,267903 Sonarworks 0,315484 Members of Cluster Number 3 (Data_nor) and Distances from Respective Cluster Center Cluster contains 6 cases Case No. Distance Case No. Distance Mapon 0,517250 eAgronom 0,377288 Sunly 0,246919 Binalyze 0,387621 Bolt 1,205217 Veriff 0,191586 Source: Author’s calculations Source: Author’s calculations 365 Organizacija, V olume 58 Issue 4, November 2025 Research Papers Table 9: Composition of the 4 cluster (STATISTICA 13 cluster analysis listing) At the same time, they have a higher-than-average amount of investment raised (Total Raised), but a smaller number of employees. This may indicate that this cluster unites startups that are possibly in a stage of active growth and development, attracting significant investment for scaling, but have not yet achieved high indicators for other criteria, such as prof- itability or media activity. Vinted, as the most distant from the cluster center, likely has characteristics that differ significantly from this typical profile, possibly demonstrating higher indicators for some criteria, which accounts for the greater distance. Table 8 demonstrates the composition of the third clus- ter obtained as a result of cluster analysis in STATISTICA 13, and the distances of the startups to the center of this cluster. Cluster 3 unites the most successful and developed startups, which aligns with Lithuania’s growing global momentum in 2025 (Baltic Tech Ventures, 2025), includes six startups: Mapon, Sunly, Bolt, eAgronom, Binalyze, and Veriff. The analysis of distances to the cluster center (Table 8) shows that Veriff (0.191586) and Sunly (0.246919) are closest to the center, indicating their high similarity to the typical characteristics of the cluster. eAgronom (0.377288) and Binalyze (0.387621) demonstrate a slightly greater distance, indicating a lesser prominence of common traits. Mapon (0.517250) is located at an even greater distance. Bolt (1.205217) is a clear outlier, significantly distant from the cluster center, which indicates its significant difference from the other group members. The startups included in Cluster 3, on average, demon- strate significantly above-average indicators for the sample across most criteria, including profitability, social media activity, investor attraction, network indicators, and media influence. They also have a higher-than-average amount of investment raised and a larger number of employees. This indicates that this cluster unites the most successful and developed startups, which have achieved significant results in all key areas. Bolt, being the most distant from the cluster center, is likely an outstanding example even within this group, possibly demonstrating extremely high values for some parameters, which accounts for its isolated position. This cluster can be characterized as a cluster of Members of Cluster Number 4 (Data_nor) and Distances from Respective Cluster Center Cluster contains 5 cases Case No. Distance Case No. Distance Tuum 0,342076 Nord Security 0,790433 BoBo 0,319088 PVcase 0,285374 Biomatter 0,269727 Source: Author’s calculations highly effective and fast-growing startups. Table 9 presents the composition of the fourth cluster obtained as a result of cluster analysis in STATISTICA 13, and the distances of the startups to the center of this cluster. Cluster 4 includes five startups: Tuum, BoBo, Bi- omatter, PVcase, and Nord Security. The analysis of distances to the cluster center (Table 9) shows that Bio- matter (0.269727) and PVcase (0.285374) are closest to the center, indicating their high similarity to the typical characteristics of the cluster. BoBo (0.319088) and Tuum (0.342076) demonstrate a slightly greater distance, indi- cating a lesser prominence of common traits. Nord Se- curity (0.790433) is significantly distant from the cluster center, which indicates its substantial difference from the other group members. The startups included in Cluster 4 are characterized, on average, by significantly below-average indicators for the sample across almost all criteria, including profitability, social media activity, investor attraction, innovativeness, social responsibility, network indicators, and media influ- ence. They also have a below-average amount of investment raised and a number of employees. This indicates that this cluster unites startups that are likely in an early stage of development or experiencing difficulties with growth and resource attraction. Nord Security, as the most distant from the cluster center, likely has characteristics that differ somewhat from this typical profile, possibly demonstrat- ing higher values for some criteria, which accounts for the greater distance. This cluster can be characterized as a cluster of nascent or struggling startups. 4.4 Typology of startups based on taxonomic analysis This subsection provides a typology of startups based on calculated taxonomic coefficients, allowing us to iden- tify groups of companies with similar characteristics. The results of calculating the taxonomy indicators are present- ed in Table 10. The visualization of the location of startups in this co- ordinate system is presented in Figure 3. 366 Organizacija, V olume 58 Issue 4, November 2025 Research Papers Table 10: Results of the taxonomic analysis of startups Startup Taxonomy coefficient 1 Factor Taxonomy coefficient 2 Factor Vinted 0,784 0,86 Mapon 0,713 0,6 Tuum 0,553 0,63 Green Genius 0,643 0,82 Origin 0,629 0,75 Sunly 0,794 0,49 BoBo 0,336 0,49 Aerones 0,612 0,81 Bolt 1,00 0,32 Ovoko 0,517 0,67 Roibox 0,587 0,81 eAgronom 0,727 0,55 Biomatter 0,346 0,55 Sonarworks 0,574 0,89 Binalyze 0,776 0,4 Nord Security 0,501 0,6 Naco 0,617 0,71 Veriff 0,77 0,52 PVcase 0,317 0,49 Cenos 0,559 0,88 where Factor 1 “Economic Potential and Structural Efficiency” combines indicators that reflect the economic sustainability and operational efficiency of startups. Variables such as expected profit (X1), investor availability (X3), technology innovativeness (X4), clustering coeffi- cient (X7), funds raised (X11), and number of employees (X13) characterize the financial strength, innovative capabilities, and structural parameters of a startup. Factor 2 “Social Engagement and Network Influence” reflects the social activity and network involvement of startups. Variables such as social media activity (X2), level of social responsibility (X5), number of connections (X6), betweenness centrality (X8), number of mentions (X9), and sentiment of mentions (X10) emphasize the importance of social reputation, audience interaction, and network influence for the success of startups. Source: Author’s calculations Figure 3 shows 4 quadrants: Quadrant I (Upper right quadrant) has “High Econom- ic Potential / High Social Engagement (HEP/HSE)”; Quadrant II (Lower right quadrant) has “High Eco- nomic Potential / Low Social Engagement (LEP/HSE)”; Quadrant III (Upper left quadrant) has “Low Economic Potential / High Social Engagement (HEP/LSE)”; Quadrant IV (Lower left quadrant) has “Low Econom- ic Potential / Low Social Engagement (LEP/LSE)”. Startup Characteristics and Recommendations. Startups in Quadrant I have a strong economic base (high profitability, investment, innovation, efficient struc - ture) and actively interact with their audience, have a de- veloped network of contacts, and a positive reputation. This is the most favorable position. Development rec- ommendations: focus on scaling the business, expanding markets, strengthening the brand, and maintaining high customer loyalty. Invest in further innovation and team development. Financing recommendations: have good op- portunities to attract both venture capital and bank loans. They can consider IPOs or M&A. Startups in Quadrant III have a strong social presence and interact well with their audience, but have not yet achieved high economic indicators. These may be young projects or projects focused on social impact rather than rapid profit. Development recommendations: need to focus on im- proving economic indicators: developing a clearer busi- ness model, searching for new sources of income, and optimizing costs. It is important to monetize the exist- ing social base. Financing recommendations: can attract grants, crowdfunding, and impact investments from inves- tors focused on social returns. It is important to demon- strate the potential for growth of the business model. 367 Organizacija, V olume 58 Issue 4, November 2025 Research Papers Startups in Quadrant II have a strong economic base but pay insufficient attention to interacting with their au- dience and building a network of contacts. There is a risk of missing opportunities for growth and development due to insufficient brand awareness and customer loyalty. De- velopment recommendations: need to actively develop so- cial networks, PR, content marketing, and participate in industry events. It is important to improve communication with clients and partners. Financing recommendations: have good opportunities to attract traditional investments (venture capital, bank loans), but it is important to show investors a plan to improve social engagement indicators. Startups in Quadrant IV are in the most vulnerable po- sition, as they have weak indicators in both economic po- tential and social engagement. Development recommen- dations: require a serious revision of the business model, searching for new ideas and development paths. It is nec- essary to improve both economic indicators and social me- dia activity. It may be necessary to involve mentors or con- sultants. Financing recommendations: attracting financing will be difficult. It may be worth considering options with bootstrapping (self-financing), grants for starting entrepre- neurs, or participation in acceleration programs. Thus, specific actions should depend on the specifics of each startup, its industry, and target market. The posi- Figure 3: Taxonomic typology matrix of startups Source: Author’s calculations tion of a startup in the matrix is not static. Companies can move from one quadrant to another as they develop. This analysis provides useful information for making strategic decisions and planning startup development. 5 Discussion The results of our study emphasize the importance of integrating analytical methods to improve the quality of investment decisions in startups, which is confirmed by a number of works. For example, the use of machine learn- ing, described in our study, is consistent with the findings of Ciampi and Gordini (2012), who note its high accuracy in forecasting defaults of small businesses. Furthermore, our observation about the significance of network analysis in assessing the market sustainability of startups is consist- ent with research by McAdam et al. (2015), which empha- sizes the importance of horizontal networks for knowledge sharing and stimulating innovation. Our findings are in line with recent studies showing the growing use of AI-driven analytics to enhance startup ecosystems and support deci- sion-making for investors (Hemanth & Lakshminarayana, 2025; Lutfiani et al., 2025). However, our analysis also revealed new aspects. For 368 Organizacija, V olume 58 Issue 4, November 2025 Research Papers example, the integration of social media analysis methods, as shown in our study, allows for taking into account repu- tational risks and public opinion in real time, which differs from traditional approaches such as expert assessments (Sivicka, 2018). This underscores the need for further study of the role of social media in investment manage- ment. Separately, it is worth noting our observation about the heterogeneity of the application of modern methods in the Baltic states. Estonia’s leadership in digital maturity mir- rors global trends where ecosystems with advanced ana- lytics outperform others (Startup Genome, 2025). While Estonia demonstrates a high level of digital maturity and actively uses analytical tools, Latvia and Lithuania re- main largely oriented towards traditional approaches. This is partially confirmed by the results of Rupeika-Apoga (2014), who notes limitations in access to modern financ- ing instruments in these countries. The contribution of our research lies in the development of a comprehensive start- up evaluation model that combines methods of taxonomy, machine learning, and network analysis. Unlike the approaches described by Fisch (2018) and Teker et al. (2016), our model allows for considering a wide range of factors, including social activity and me- dia influence, which is particularly relevant for startups focused on long-term growth. Thus, the results confirm the significance of the proposed methodology and open up prospects for its further application in other regions and industries. Moreover, the integration of sustainability and digitalization in startup evaluation is emphasized in the Green Startup Report 2025 (Fichter et al., 2025), which highlights the potential of digital tools for supporting green innovation. However, further research could focus on assessing the long-term effectiveness of the proposed model in a changing business environment. 6 Conclusions The application of modern business analytics methods, such as machine learning, network analysis, and social media analysis, allows for increased accuracy in assessing the prospects of startups. These methods demonstrate high efficiency: for example, the use of machine learning allows achieving default prediction accuracy of 98.6% (Ciampi and Gordini, 2012), and network analysis identifies key players and relationships with centrality coefficients up to 0.995. How do modern analytical methods compare to tradi- tional approaches in startup valuation? The results demonstrate that hybrid models combin- ing financial metrics with digital indicators (e.g., social media activity, network centrality) outperform traditional methods (e.g., SWOT, expert assessments), reducing sub- jectivity and improving accuracy.Which factors (financial, social, technological) are most critical for startup success in the Baltics? Factor analysis revealed that economic potential (prof- itability, investor availability) and social engagement (on- line activity, sentiment) are the primary drivers, explaining 88.5% of the variance in startup rankings. The findings strongly support the hypothesis (H1) that data-driven methods enhance decision-making accuracy, as evidenced by the high correlation coefficients (> 0.98) for key variables. Factor analysis revealed that economic potential (profitability, investor availability) and social en- gagement (online activity, sentiment) are the primary driv- ers, explaining 88.5% of the variance in startup rankings. The developed startup evaluation model, which integrates taxonomy, machine learning, and social media analysis, outperforms traditional approaches by reducing subjectiv- ity and improving reliability. The key factors determining the success of startups in the Baltic states are economic stability, technological in- novativeness, social activity, and media influence. Factor analysis showed that financial and resource potential (fac- tor loading coefficient 0.986) and the level of social media engagement (coefficient 0.990) have the highest correla- tion with startup success. A comparative analysis of the Baltic countries revealed significant differences in startup assessment approaches. In Estonia, modern methods account for 45% of the total number of approaches used, including machine learning al- gorithms and Monte Carlo simulations, which underscores its leadership in digitalization. Latvia and Lithuania use traditional methods in 60% and 50% of cases, respective- ly, which limits their competitiveness in the global startup ecosystem. Taxonomic and cluster analysis made it possi- ble to identify groups of startups with different levels of economic and social potential. Companies with high eco- nomic stability (average taxonomy coefficient 0.784) and social activity (average coefficient 0.86) occupy leading positions. Conversely, startups with low indicators, such as companies with a taxonomy coefficient below 0.5 (Bi- omatter, BoBo), need to revise their business models and require support. The developed startup evaluation model, which integrates taxonomy, machine learning, and social media analysis, has proven its applicability for investment decision-making and can be adapted for other regions and sectors of the economy. For successful development, start- ups are recommended to focus on strengthening financial stability, increasing social engagement, and enhancing their reputation, while investors are advised to integrate analytical tools into decision-making to minimize risks and increase returns. Theoretical implications. This study contributes to the literature on startup evaluation by proposing a hybrid multifactor framework that integrates financial, techno- logical, and social indicators. It extends prior research by demonstrating the value of combining traditional financial 369 Organizacija, V olume 58 Issue 4, November 2025 Research Papers metrics with digital signals such as network centrality and social media activity in a unified model. Practical implications. The results provide investors with evidence-based tools for more accurate and timely startup evaluation, helping to reduce risks and improve de- cision-making quality. Policymakers and startup support organizations can also use the findings to design programs that strengthen financial stability, foster social engage- ment, and encourage the adoption of advanced analytics in the Baltic startup ecosystem. Limitations. The study is limited to startups in the Bal- tic States and relies on a sample of 20 companies, which may affect the generalizability of the results. In addition, the analysis is based on historical data and selected indica- tors, so incorporating a broader range of variables or lon- gitudinal data could provide deeper insights. Future research. Further studies should explore how the resilience and transformation of Baltic startups (LSM, 2025; Stats and Market Insights, 2025) will shape long-term investment strategies. Expanding the dataset to include other regions and additional indicators—such as ESG metrics or customer sentiment—could further vali- date and enhance the proposed model. Acknowledgements This work of O. Dorokhov and K. Ukrainski was sup- ported by the project „Increasing the knowledge intensity of Ida-Viru entrepreneurship“ co-funded by the European Union. References Anuradha, A., & Sailaxmi, S. S. (2024). Growth hacking business analytical tools for small and medium enter- prises. Journal of Business Analytics, 7(4), 292–317. https://doi.org/10.1080/2573234x.2024.2387560 Acebo, E., Miguel-Dávila, J., & Nieto, M. (2020). Do fi- nancial constraints moderate the relationship between innovation subsidies and firms’ R&D investment? European Journal of Innovation Management, 25(2), 347–364. https://doi.org/10.1108/ejim-07-2020-0286 Aernoudt, R. (2004). Incubators: a tool for entrepreneur- ship? Small Business Economics, 23(2), 127–135. https://doi.org/10.1023/b:sbej.0000027665.54173.23. Agrawal, A., Catalini, C., & Goldfarb, A. (2014). Some simple economics of crowdfunding. Innovation Policy and the Economy, 14, 63–97. https://doi. org/10.1086/674021 Al-Debei, M. (2023). The era of business analytics: iden- tifying and ranking the differences between business intelligence and data science from practitioners’ per- spective using the Delphi method. Journal of Business Analytics, 7(2), 94–119. https://doi.org/10.1080/25732 34x.2023.2285483 Anstead, N., & O’Loughlin, B. (2014). Social media anal- ysis and public opinion: The 2010 UK General Elec- tion. Journal of Computer-Mediated Communication, 20(2), 204–220. https://doi.org/10.1111/jcc4.12102 Baijens, J., Huygh, T., & Helms, R. (2021). Establishing and theorising data analytics governance: a descriptive framework and a VSM-based view. Journal of Busi- ness Analytics, 5(1), 101–122. https://doi.org/10.1080/ 2573234x.2021.1955021 Baltic Tech Ventures. (2025, January 20). Lithuania’ s startup ecosystem: Gaining global momentum in 2025. Baltic Tech Ventures. https://www.baltictechventures. com/post/lithuania-s-startup-ecosystem-gaining-glob- al-momentum-in-2025 Berlemann, M., & Jahn, V . (2015). Regional importance of Mittelstand firms and innovation performance. Re- gional Studies, 50(11), 1819–1833. https://doi.org/10. 1080/00343404.2015.1058923 Bigliardi, B., Dolci, V ., Filippelli, S., Petroni, A., & Pini, B. (2025). The role of digitalization in startups: a keyword-based literature review. Procedia Computer Science, 253, 2665–2673. https://doi.org/10.1016/j. procs.2025.01.326 Bizri, R. M. (2017). Refugee-entrepreneurship: a social capital perspective. Entrepreneurship and Regional Development, 29(9–10), 847–868. https://doi.org/10.1 080/08985626.2017.1364787 Blanquet, L. B., Pereira, M. A., & Petrov, S. (2025). An interpretable machine learning framework for explain- ing company valuation. Decision Analytics Journal, 100611. https://doi.org/10.1016/j.dajour.2025.100611 Brecht, P., Niever, M., Kerres, R., Ströbele, A., & Hahn, C. H. (2021). Smart platform experiment cycle (SPEC): a process to design, analyze, and validate digital plat- forms. Artificial Intelligence for Engineering Design Analysis and Manufacturing, 35(2), 209–225. https:// doi.org/10.1017/s0890060421000081 Calabrese, R., & Osmetti, S. (2013). Modelling small and medium enterprise loan defaults as rare events: the ge- neralized extreme value regression model. Journal of Applied Statistics, 40(6), 1172–1188. https://doi.org/1 0.1080/02664763.2013.784894 Chatterji, A., Delecourt, S., Hasan, S., & Koning, R. (2018). When does advice impact startup performan- ce? Strategic Management Journal, 40(3), 331–356. https://doi.org/10.1002/smj.2987 Ciampi, F., & Gordini, N. (2012). Small Enterprise Default Prediction Modeling through Artificial Neural Networ- ks: An Empirical Analysis of Italian Small Enterprises. Journal of Small Business Management, 51(1), 23–45. https://doi.org/10.1111/j.1540-627x.2012.00376.x Coleman, S., Cotei, C. and Farhat, J. (2016), «The debt-equity financing decisions of U.S. startup firms», 370 Organizacija, V olume 58 Issue 4, November 2025 Research Papers Journal of Economics and Finance, V ol. 40 No. 1, pp. 105-126. Dada, O., Jack, S., & George, M. (2015). University– Business Engagement Franchising and Geographic Distance: A case study of a Business leadership pro- gramme. Regional Studies, 50(7), 1217–1231. https:// doi.org/10.1080/00343404.2014.995614 Dealroom database - Everyone is here - Startup Lithuania. (2022, November 29). Startup Lithuania. https://www. startuplithuania.com/dealroom-database/ EU-Startups. (2023, January). Booming Baltics: Trends and predictions for 2023. Retrieved from https://www. eu-startups.com/2023/01/booming-baltics-trends-and- predictions-for-2023/ EU-Startups. (2023, March). 10 super promising startups from Latvia to keep an eye on in 2023 and beyond. Re- trieved from https://www.eu-startups.com/2023/03/10- super-promising-startups-from-latvia-to-keep-an-eye- on-in-2023-and-beyond/ Fichter, K., Neumann, T., Olteanu, Y ., & Grothey, T. (2025, April). Green startup report 2025. Research- Gate. https://doi.org/10.13140/RG.2.2.29610.32968 Fisch, C. (2018). Initial coin offerings (ICOs) to finance new ventures. Journal of Business Venturing, 34(1), 1–22. https://doi.org/10.1016/j.jbusvent.2018.09.007 Foreman-Peck, J., Makepeace, G., & Morgan, B. (2006). Growth and profitability of small and me- dium-sized enterprises: Some Welsh evidence. Regional Studies, 40(4), 307–319. https://doi. org/10.1080/00343400600725160 Foster, E. (2004). Research on Gossip: Taxonomy, meth- ods, and future directions. Review of General Psy- chology, 8(2), 78–99. https://doi.org/10.1037/1089- 2680.8.2.78 Gazzola, P., Paterson, A., Amelio, S., & Ferioli, M. (2022). Certified B Corporations and Innovation: Crowdfund- ing as a tool for Sustainability. Sustainability, 14(24), 16639. https://doi.org/10.3390/su142416639 Hemanth, J. B., & Lakshminarayana, K. (2025, April). AI and the evolution of IT startup ecosystems: A global perspective. International Journal of Scientific Re- search in Engineering and Management, 8(1). https:// doi.org/10.55041/IJSREM44391 Ioakeimidou, D., Chatzoudes, D., & Chatzoglou, P. (2024). Assessing data analytics maturity: propos- ing a new measurement scale. Journal of Business Analytics, 1–15. https://doi.org/10.1080/257323 4x.2024.2439990 Jabeur, S., & Fahmi, Y . (2017). Forecasting financial dis- tress for French firms: a comparative study. Empirical Economics, 54(3), 1173–1186. https://doi.org/10.1007/ s00181-017-1246-1 Kato, T., Kamei, S., Ootsubo, T., & Ichiki, Y . (2023). More information is not better: examining appropriate information for estimating sales performance in con- cept testing. Journal of Business Analytics, 6(3), 188– 202. https://doi.org/10.1080/2573234x.2023.2167670 Kaya, M. C., & Persson, L. (2019). A Theory of Gazelle Growth: Competition, Venture capital finance and pol- icy. SSRN Electronic Journal. https://doi.org/10.2139/ ssrn.3680711 Keogh, D., & Johnson, D. (2021). Survival of the funded: Econometric analysis of startup longevity and success. Journal of Entrepreneurship Management and Innova- tion, 17(4), 29–49. https://doi.org/10.7341/20211742. Kim, M., Han, Y . S., & Cho, R. (2024). The role of co- operative R&D in innovation performance of SMEs: evidence from South Korean materials, parts, and equipment firms. Asian Journal of Technology Inno- vation, 1–26. https://doi.org/10.1080/19761597.2024. 2349210 Kolodiziev, O., Gukaliuk, A., Shcherbak, V ., Riabovolyk, T., Androshchuk, I., & Pas, Y . (2024). The impact of refugee startups on host country economies: business models and economic adaptation. Economic Studies journal, Bulgarian Academy of Sciences - Econom- ic Research Institute, 2, 175-201. https://ideas.repec. org/a/bas/econst/y2024i2p175-201.html Langfelder, P., & Horvath, S. (2008). WGCNA: an R pack- age for weighted correlation network analysis. BMC Bioinformatics, 9(1). https://doi.org/10.1186/1471- 2105-9-559 Liu, Y ., Zeng, Q., Li, B., Ma, L., & Ordieres‐Meré, J. (2022). Anticipating financial distress of high‐tech startups in the European Union: A machine learning approach for imbalanced samples. Journal of Fore- casting, 41(6), 1131–1155. https://doi.org/10.1002/ for.2852 LSM. (2025, February 1). 2025 predictions for the Lat- vian startup scene. LSM.lv. https://eng.lsm.lv/article/ economy/business/01.02.2025-2025-predictions-for- the-latvian-startup-scene.a586010/ Lu, Y . (2019). Bayesian assessment of predictors’ contri- butions to variation in the predictive performance of a logistic regression model. Journal of Business An- alytics, 2(2), 134–146. https://doi.org/10.1080/257323 4x.2019.1678400 Lutfiani, N., Wijono, S., Rahardja, U., & Purnomo, H. D. (2025). Advancing startup ecosystems through AI-Driven Matchmaking: A comprehensive biblio- metric analysis. International Journal of Engineering Science and Information Technology, 5(1), 446–454. https://doi.org/10.52088/ijesty.v5i1.1095 McAdam, M., McAdam, R., Dunn, A., & McCall, C. (2015). Regional Horizontal Networks within the SME Agri-Food Sector: An Innovation and Social Network Perspective. Regional Studies, 50(8), 1316–1329. https://doi.org/10.1080/00343404.2015.1007935 Melegati, J., Goldman, A., Kon, F., & Wang, X. (2019). A 371 Organizacija, V olume 58 Issue 4, November 2025 Research Papers model of requirements engineering in software start- ups. Information and Software Technology, 109, 92– 107. https://doi.org/10.1016/j.infsof.2019.02.001 Murphy, K. (2012). Machine Learning : A Probabilistic Perspective. http://cds.cern.ch/record/1981503 O’Neill, M., & Brabazon, A. (2019). Business analytics capability, organisational value and competitive ad- vantage. Journal of Business Analytics, 2(2), 160–173. https://doi.org/10.1080/2573234x.2019.1649991 Palacín-Sánchez, M., & Di Pietro, F. (2015). The role of the regional financial sector in the capital structure of Small and Medium-Sized Enterprises (SMEs). Region- al Studies, 50(7), 1232–1247. https://doi.org/10.1080/ 00343404.2014.1000290 Qin, H., Koong, K., Wen, H., & Liu, L. (2022). Mapping business analytics skillsets with industries: empirical evidence from online job advertisements. Journal of Business Analytics, 6(3), 167–179. https://doi.org/10.1 080/2573234x.2022.2136541 Quansah, E. (2024). Digitalization the necessary evil: in- tegrating digital technologies in businesses of BOP countries. Information Technology for Development, 1–29. https://doi.org/10.1080/02681102.2024.243289 0 Rupeika-Apoga, R. (2014). Alternative financing of SMEs in the Baltic States: myth or reality? Procedia - Social and Behavioral Sciences, 156, 513–517. https://doi. org/10.1016/j.sbspro.2014.11.231 Sivicka, J. (2018). Features of valuation of startup com- panies. Economic Scope, 0(132), 163–174. https://doi. org/10.30838/p.es.2224.240418.163.60 Startup Estonia. (2023). Annual Report: Estonia’s Startup Ecosystem. Retrieved from https://www.startupesto- nia.ee/annual-report Startup Genome. (2025). The Global Startup Ecosys- tem Report 2025. https://startupgenome.com/report/ gser2025/introduction Startup Lithuania. (n.d.). Startup Lithuania. https://www. startuplithuania.com/ Stats and Market Insights. (2025b, January 15). Lith- uania startup ecosystem in 2025: A year of resil- ience and transformation. Stats and Market Insights. https://www.statsandmarketinsights.com/blog/85/ lithuania-startup-ecosystem-in-2025-a-year-of-resil- ience-and-transformation Stats and Market Insights. (2025а, January 15). Latvia startup ecosystem in 2025: A year of resilience and transformation. Stats and Market Insights. https:// www.statsandmarketinsights.com/blog/36/latvia-start- up-ecosystem-in-2025-a-year-of-resilience-and-trans- formation Sutherland, M., Thompson, D., & Edirisingha, P. (2019). Try before you buy: a small business employer (SME) perspective of international student mobility in Eng- land. Studies in Higher Education, 46(7), 1256–1271. https://doi.org/10.1080/03075079.2019.1680965 Teker, D., Teker, S., & Teraman, Ö. (2016). Venture Capital Markets: A Cross Country analysis. Procedia Econom- ics and Finance, 38, 213–218. https://doi.org/10.1016/ s2212-5671(16)30192-7 Tomczak, A., & Brem, A. (2013). A conceptualized invest- ment model of crowdfunding. Venture Capital, 15(4), 335–359. https://doi.org/10.1080/13691066.2013.847 614 Wahlstrøm, R., Becker, L., & Fornes, T. (2024). Enhanc- ing credit risk assessments of SMEs with non-finan- cial information. Cogent Economics & Finance, 12(1). https://doi.org/10.1080/23322039.2024.2418910 Wong, P., Ho, Y ., & Autio, E. (2005). Entrepreneurship, In- novation and Economic Growth: Evidence from GEM data. Small Business Economics, 24(3), 335–350. https://doi.org/10.1007/s11187-005-2000-1 Wright, M., Westhead, P., & Ucbasaran, D. (2007). Inter- nationalization of Small and Medium-sized Enterprises (SMEs) and International entrepreneurship: a critique and policy implications. Regional Studies, 41(7), 1013– 1030. https://doi.org/10.1080/00343400601120288 Yaakobi, A., Goresh, M., Reychav, I., McHaney, R., Zhu, L., Sapoznikov, H., & Lib, Y . (2019). Organisational project evaluation via machine learning techniques: an exploration. Journal of Business Analytics, 2(2), 147– 159. https://doi.org/10.1080/2573234x.2019.1675478 Valeriia Shcherbak is a Professor at the Department of Economic and Entrepreneurship, Sumy National Agrarian University, Ukraine, and also works at the Poltava University of Economics and Trade, Ukraine. She earned her Doctor of Economic Sciences degree in 2009 and received the academic title of Professor at V. N. Karazin Kharkiv National University, Ukraine. Her research focuses on sustainable rural development, digital transformation, inclusive economy, geoinformation platforms for tourism, and refugee integration during crises. She is the author of the textbook Marketing Distribution Policy and has published more than 300 scientific articles. Oleksandr Dorokhov is a Visiting Professor at the Department of Public Economics and Policy, University of Tartu, Estonia. He earned his PhD in Technical Science from the Kharkiv National Automobile and Highway University, Ukraine. Until February 2022, he worked as a Professor of the Department of Information Systems at the Simon Kuznets Kharkiv National University of Economics, Ukraine. His research focuses on multicriteria decision support systems, computer modeling in economics, fuzzy logic, and modeling the functioning of startups and entrepreneurial ecosystems. 372 Organizacija, V olume 58 Issue 4, November 2025 Research Papers Kadri Ukrainski is a Professor in Research and Innovation Policy and Head of the Department of Public Economics and Policy, University of Tartu, Estonia. She earned her PhD from the Faculty of Economics and Business Administration at the University of Tartu, Estonia. Her research focuses on science and innovation policy, startups, high/deep-tech technologies, and the possibilities that public sector agents, such as universities, ministries, SOEs, and agencies, have to facilitate such processes in societies. Deniss Djakons is a Professor and Rector at ISMA University of Applied Sciences, Latvia. He earned his Dr.oec (Doctor of Economics) degree and currently leads the institution while maintaining an active research profile. His scholarly work focuses on higher education financing, strategic management of territorial development, and the social responsibility of businesses in global supply chains. His publications particularly examine innovation policy, university development strategies, and comparative studies of education systems in transition economies. Olha Kovalyova is an Associate Professor at the Department of Economics and Entrepreneurship, Sumy National Agrarian University, Ukraine. She earned her Ph.D. in Economics and specializes in agricultural economics and sustainable land management. Her research focuses on optimizing economic potential in agro-industrial sectors, sustainable fertilizer management, and innovative approaches to agricultural enterprise development. She has contributed significantly to studies on farmland evaluation, food security, and ecological-economic systems in agriculture. Liudmyla Dorokhova is a Visiting Professor at the Department of Marketing, University of Tartu, Estonia. She earned her PhD in Pharmacy from the National University of Pharmacy, Ukraine. Until February 2022, he worked as a Professor at the Department of Marketing, National University of Pharmacy, Ukraine. Her research focuses on medical and healthcare startups, consumer behavior and choice, and decision- making models. 373 Organizacija, V olume 58 Issue 4, November 2025 Research Papers Appendix: Data about startups Source: https://www.seedtable.com/best-startups-in-lithuania; https://www.seedtable.com/best-startups-in-latvia; https://www.seedtable.com/best-startups-in-estonia Startup Country Projected profitability, M$ Activity in social net- works, ths. subscribers Avail- ability of investors, number Innovativeness of technologies, scores 1-10 Basic level of social respon- sibility, score 1-10 Number of links, node degree Cluster coefficient Betweenness centrality Number of mentions Senti- ment of mentions Total Raised, M$ Total Raised, M$ Number of em- ployees, ths. of people Х1 Х2 Х3 Х4 Х5 Х6 Х7 Х8 Х9 Х10 Х11 Х12 Х13 Vinted Lithuania 2 5 2 8 7 10 0,60 50 100 60 677,54 12 1 Mapon Latvia 5 10 5 9 9 15 0,80 100 200 150 3 2996 0,1 Tuum Estonia 1 2 1 6 6 5 0,40 20 50 20 49,82 202 0,1 Green Genius Lithuania 3 7 3 7 8 12 0,70 70 150 100 109,8 193 0,2 Origin Latvia 4 9 4 8 7 14 0,75 90 180 120 4,35 3478 0,2 Sunly Estonia 6 12 6 9 8 18 0,85 120 250 180 364,84 206 0,2 BoBo Lithuania 0,5 1 0 5 5 3 0,20 5 20 -10 7,01 803 0,01 Aerones Latvia 2,5 6 2 7 6 9 0,55 40 90 50 12,01 4617 0,25 Bolt Estonia 7 15 7 10 9 20 0,90 150 300 220 1015,71 45 5 Ovoko Lithuania 1,5 3 1 6 7 7 0,45 30 70 30 22,2 917 0,2 Roibox Latvia 3,5 8 3 8 8 13 0,72 80 160 110 3,29 56,01 0,05 eAgronom Estonia 5,5 11 5 9 9 17 0,82 110 230 160 13,19 704 0,5 Biomatter Lithuania 0,8 1,5 0 4 6 4 0,30 10 30 -5 7,2 1144 0,05 Sonar- works Latvia 2,8 6,5 2 7 7 11 0,65 60 110 70 5,6 6886 0,01 Binalyze Estonia 6,5 13 6 9 8 19 0,88 130 270 190 30,81 553 0,1 Nord Security Lithuania 1,2 2,5 1 5 6 6 0,35 25 60 15 100 1144 5 Naco Latvia 4,2 9,5 4 8 7 15 0,78 95 190 130 1,65 6886 0,01 Veriff Estonia 5,8 11,5 5 9 8 18 0,84 115 240 170 184,62 639 1 PVcase Lithuania 0,7 1,2 0 4 5 3 0,25 8 25 -8 100,35 1452 0,25 Cenos Latvia 3,2 7,5 3 7 6 12 0,68 65 130 80 1,47 7742 0,05