Unpacking the performance of corporate accelerators: can management structure and strategic focus explain their success in supporting startup exits?

Corporate accelerators represent a recent fast-emerging corporate venturing mode that large corporations can leverage to reach their strategic objectives and grasp entrepreneurial opportunities by supporting startups and catalyzing their growth (Kohler, 2016; Kupp et al., 2017; Urbaniec and Żur, 2021). Through fixed-term, cohort-based programs, corporate accelerators provide startups with critical resources (e.g. mentorship, coworking spaces, networking, and funding) to let them survive and scale, potentially become valuable partners for corporations, and even make exits (Cohen et al., 2019).

As largely recognized, the landscape of corporate accelerators exhibits considerable variation, and acceleration programs differ across multiple dimensions, including, among others, their location, size, services, design features (e.g. cohort size, investment models, and length of program), management structure, strategic focus, and founding sponsors (Cohen et al., 2019; Moschner et al., 2019; Shankar and Shepherd, 2019). So far, scant research has explored the corporate accelerators’ landscape, exploring the relations between corporate accelerators’ dimensions and performance (i.e. Canovas-Saiz et al., 2020, 2021). These few studies have suggested that the length of acceleration programs, the geographical location, and cohort size affect the accelerators’ success in supporting startups to survive, receive funding from different investors, and exit via an Initial Public Offering (IPO) (Canovas-Saiz et al., 2020, 2021). Thus, the scarcity of these studies highlights a gap that calls for more in-depth investigations to enhance our understanding of accelerators’ performance. In this paper, we take one of the possible research directions aimed at filling this literature gap by considering further dimensions that characterize corporate accelerators for unpacking their performance. Specifically, the objective of our study is to explore whether and how their management structure and strategic focus drive corporate accelerators’ performance. While prior research has recognized these two dimensions as critical in the corporate accelerators’ landscape (Moschner et al., 2019; Seitz et al., 2024; Shankar and Shepherd, 2019), their role as drivers of performance has remained largely overlooked. By addressing this gap, our study contributes novel insights into the determinants of corporate accelerators’ effectiveness. Considering the management structure, corporations running an accelerator can either decide to internally organize the activities of their programs (i.e. in-house accelerators) or delegate the organization of such programs to independent external organizations (i.e. powered-by accelerators) (Moschner et al., 2019). As regards the strategic focus, corporations can either run acceleration programs that focus exclusively on those startups operating in industries fitting with their core business (i.e. vertical accelerators) or programs that have a very broad or no focus by admitting startups with diverse industrial backgrounds (i.e. horizontal accelerators) (Seitz et al., 2024).

Shedding new light on the relationship between unexplored dimensions of corporate accelerators and their performance holds significant relevance for both academia and startup communities. Investigating unexplored facets of corporate accelerators and their impact on performance provides scholars with new evidence that deepens the understanding of such a novel corporate venturing mode. Furthermore, by uncovering critical insights into the dynamics between these unexplored dimensions and accelerator performance, the research offers valuable information for corporate managers and entrepreneurs. Particularly, this information can help corporate managers to better design their acceleration programs and entrepreneurs to alleviate the uncertainties faced by startups when selecting an acceleration program, enabling them to make more informed decisions.

To reach our aim, we gathered data considering 188 corporate accelerators registered until February 2023 on the Crunchbase database. Then, we performed an econometric analysis to assess whether and how differences in the corporate accelerators’ performance can be explained by their management structure and strategic focus. Our results suggest that in-house corporate accelerators are more effective in supporting startups’ exits (via IPO or acquisition) compared to powered-by accelerators. We also found that accelerators closely aligned with a corporation’s core business (vertical) are less effective at driving exits than those with a broader focus (horizontal). Furthermore, in-house accelerators with a broader or no strategic focus (horizontal) perform better in supporting exits than vertical accelerators. Regarding startup survival, in-house accelerators are more successful than powered-by ones, and vertical accelerators are better at supporting survival than horizontal ones. However, horizontal in-house accelerators outperform vertical ones in facilitating exits.

The results of this paper offer important theoretical and practical implications. Particularly, this paper adds new evidence to the few recent studies that have investigated the corporate accelerators’ performance (Canovas-Saiz et al., 2020, 2021) by suggesting that the management structure and the strategic focus represent two key corporate accelerators’ attributes that can predict their success. Moreover, this study offers to the existing literature investigating the effect of corporate venturing modes on the startups’ exiting (Ivanov and Xie, 2010; Kim and Haemin, 2017; Wang et al., 2022) by highlighting how, beyond more traditional corporate venturing modes (e.g. CVC investments, acceleration programs can affect the startups’ likelihood of going public or being acquired too. Furthermore, this study provides valuable practical recommendations for entrepreneurs assessing corporate accelerator programs, advocating for a selection approach that aligns with their strategic objectives. The choice of an accelerator goes beyond securing industry support—it involves identifying an ecosystem that best meets the startup’s specific needs, whether for survival, scaling, or a successful market exit. Therefore, entrepreneurs should carefully evaluate corporate accelerators’ management structure and strategic focus, as these elements significantly impact access to resources, networking opportunities, and the startup’s long-term trajectory. Finally, our study offers recommendations for corporate managers, highlighting the importance of designing acceleration programs that foster accelerated startups’ exits. By adopting appropriate management structures and focusing on strategic focuses that facilitate startups’ pathways to IPOs and acquisitions, corporations can enhance their value proposition and attract high-potential startups to achieve their strategic goals.

Although their common objective is to support entrepreneurs in developing their startups and help these startups to survive, grow, and succeed, corporate accelerators substantially vary across various dimensions, with prior literature indicating that the management structure and strategic focus stand out as particularly relevant to describe their different models (Moschner et al., 2019; Seitz et al., 2024; Shankar and Shepherd, 2019). In the following, we describe these two dimensions along with their main benefits.

Considering the management structure, corporations running an accelerator can either decide to internally organize the activities of the acceleration program or delegate the organization of such a program to an independent external organization (Moschner et al., 2019).

Corporate accelerators internally managed by corporations are referred to as “in-house” since they are integrated into the corporation’s structure. Many large corporations, such as Novartis and Telefonica, use this model to sense entrepreneurial opportunities and support innovative startups. In-house corporate accelerators present several advantages that can significantly impact the survival, growth trajectory, and exit of the participating startups. Firstly, in-house accelerators provide startups with tailored and adaptive support (Moschner et al., 2019). Indeed, because an internal management structure provides full control over the accelerator, corporations can better and more flexibly adapt the acceleration programs to the startups’ progress and needs (Kanbach and Stephan, 2016). Such an inherent adaptability ensures that the support provided by in-house corporate accelerators remains relevant and responsive to the evolving challenges and opportunities faced by the startups, thus enhancing the likelihood of these startups surviving and succeeding (Möllmann, 2023). Furthermore, in-house accelerators offer startups more direct access to the corporation’s extensive knowledge and resources (Moschner et al., 2019). Since in-house accelerators are generally located within or close to the headquarters of the corporations, they allow corporations and startups to establish trustworthy and strict cooperation that activates learning processes, permits knowledge and culture sharing, and facilitates resource exchange (Holdt Christensen and Pedersen, 2018). Direct access to corporations and the tight cooperation enabled by in-house corporate accelerators allow startups to more easily tap into the knowledge and experience of seasoned professionals, leverage existing infrastructure, and access market insights from corporations (Möllmann, 2023; Moschner et al., 2019), which can greatly boost the advancement of startups.

On the other hand, corporate accelerators whose activities are run by independent external organizations (e.g. independent accelerators, venture capital firms, and innovation hubs) on behalf of corporations are known as “powered-by”. This corporate accelerator model envisions that independent external organizations that have widespread experience in running acceleration programs and supporting the startups’ growth offer facilities and professional assistance, while parent corporations usually provide financial resources (Moschner et al., 2019). Common examples are represented by the Metro Accelerator powered by Techstars, and the BNP Paribas Accelerator powered by Plug and Play. Powered-by accelerators offer numerous benefits that can have a substantial influence on the growth path of startups and their likelihood to exit. First, since external independent organizations are specialists in running acceleration programs, they can bring a wealth of expertise in evaluating startups’ business models, market potential, and growth strategies, which corporations may often lack (Moschner et al., 2019). Moreover, startups can benefit from participating in powered-by corporate accelerators since external independent organizations typically have an extensive network of relationships within the startup ecosystem that spreads largely beyond the parent corporations of the acceleration program (Kohler, 2016). By accessing knowledge from different actors in the ecosystems, startups avoid being limited to shaping their strategy and product to align exclusively with the goals and offerings of the corporations sponsoring the acceleration programs (Kohler, 2016). Differently, tapping the wide range of relationships of the independent external organization, startups can gain access to diverse perspectives and ideas and acquire a more comprehensive range of skills and expertise, which represent crucial conditions for fostering innovation and addressing the multifaceted challenges that arise during their growth (March, 1991). Additionally, the external independent organizations running powered-by corporate accelerators can support startups by exposing them to funding opportunities from a wide range of prominent investors, such as venture capitalists (Ester, 2017). Relying on the reputation of the external independent organizations, which are recognized as very skillful in screening and selecting high-profile startups, investors may likely fund those startups participating in their powered-by acceleration programs (Dalle et al., 2023). Leveraging the reputation of the professional external independent organization that runs the acceleration program on behalf of the corporation can thus offer decisive support to startups by increasing their image and credibility, boosting their fundraising opportunities, and improving their chance to survive and succeed (Woolley and MacGregor, 2022).

When designing their acceleration programs, corporations can either emphasize an exploitative strategic focus by admitting startups with an industrial background fitting their core businesses or an explorative one by accepting applications from startups across a variety of industries (Schildt et al., 2005; Shankar and Shepherd, 2019).

Corporate accelerators aiming at exploiting startups’ innovations closely related to the corporations’ existing operations and product portfolios are recognized as “vertical” or “specialist” (Kohler, 2016; Prexl et al., 2019). The Disney Accelerator represents an example of a vertical accelerator that focuses on exploiting startups’ knowledge and technologies directly relevant to the corporation’s core businesses. Indeed, its acceleration program primarily seeks startups operating within industries that are relevant to Disney, such as media and entertainment. Vertical corporate accelerators offer various benefits that can profoundly influence the survival, growth path, and likelihood of exit of the startups involved in their programs. First, in a vertical accelerator, startups can harness potential synergies with the corporation due to the interconnectedness of their core businesses (Seitz et al., 2024). Since operating in the same industry, the corporation and the startups possess similar cognitive structures, which enable and facilitate technical learning processes and knowledge transfer (Kogut and Zander, 1992). Particularly, within a vertical corporate accelerator, startups can leverage their absorptive capacity to understand and integrate specialized technical expertise and industry-related knowledge they may need to successfully develop and commercialize their innovations (Cohen and Levinthal, 1990; Kohler, 2016; Shuwaikh and Dubocage, 2022). Moreover, by participating in vertical acceleration programs, startups receive mentorship assistance from corporations that, since representing prominent clients, can support startups’ growth and performance (Gimmy et al., 2017; Moschner et al., 2019). Vertical accelerators can, for instance, assist startups in conducting pilot projects by granting them access to essential resources possessed by the parent corporations, such as production facilities and testing laboratories, for validating their products in a real-world setting (Kohler, 2016; Kurpjuweit and Wagner, 2020). Finally, the fit between the industrial background of the corporation and that of the startups in the acceleration program allows corporations to have a better understanding of the startups’ technologies, a condition that may help them to better nurture the startups’ innovation and let them access more funding and growth (Ivanov and Xie, 2010; Seitz et al., 2024).

On the other hand, corporate accelerators with an exploratory strategic approach, welcoming applications from startups spanning different industries and with diverse backgrounds, are known as “horizontal” or “generalist” (Kohler, 2016; Prexl et al., 2019). The acceleration program of Google (i.e. Google for Startups) exemplifies a horizontal accelerator model. This program focuses on supporting startups operating across various industries, united solely by their commitment to developing innovative technologies to solve the world’s important challenges and address the needs of consumers, businesses, and society. For instance, alumni of Google for Startups include startups operating in the healthcare, semiconductor, and energy industries beyond the digital one. Horizontal corporate accelerators provide a range of advantages that can significantly impact the trajectory of growth and the performance of the startups participating in their programs. First, within horizontal corporate accelerators, startups can access heterogeneous knowledge and skills both from the corporations running the acceleration programs and from startups in their cohort, which are characterized by different backgrounds (Seitz et al., 2024). Such heterogeneity provides startups with the possibility to recombine diverse pieces of knowledge and different skills, improving their ability to develop innovations (Lee et al., 2019; March, 1991). Consequently, accessing and recombining diverse knowledge in a horizontal acceleration program holds the promise of boosting the startups’ growth and increasing their likelihood of success (Belderbos et al., 2018). Furthermore, by maintaining a horizontal focus, these acceleration programs avoid restricting startups to a specific industry, preventing the lock-in effect, i.e. the startups’ risk of losing entrepreneurial freedom when sharing the same background as a partner corporation due to potential dependencies and limited diversification (Haslanger et al., 2023; Noviaristanti et al., 2024). For example, in the context of corporate accelerators, the similarity between the background of the corporations and that of the startups can restrict startups from producing products or services to satisfy specific corporations’ needs. This circumstance can lead to a constrained innovation environment, hindering the startups’ survival and success by limiting their ability to explore unconventional paths and maintain agility in decision-making (Assenova and Raphael, 2024; March, 1991). On the contrary, by ensuring that startups retain their entrepreneurial freedom while benefiting from corporations’ support and resources, horizontal corporate accelerators promote startups’ potential by encouraging innovation across various sectors (Seitz et al., 2024).

We retrieved secondary data from the Crunchbase database, which is a prominent data source including information about startups and different kinds of investors (e.g. accelerators and venture capitalists) that has been largely used by previous scholars addressing the accelerators phenomenon (Canovas-Saiz et al., 2021; Yu, 2020). Particularly, we collected data considering 188 corporate accelerators registered on Crunchbase until February 2023.

Then, the corporate accelerators’ websites have been screened to classify the corporate accelerators according to their management structure and strategic focus. In detail, two graduate assistants navigated the websites of the corporate accelerators to check whether the management structure of the corporate accelerators is internal (i.e. in-house) or delegated to a third party (i.e. powered-by). Similarly, the two graduate assistants checked for the corporate accelerators’ strategic focus assessing whether the programs hinge exclusively on industries related to the core business of the corporations (i.e. vertical) or they are dedicated to a broader set of startups’ industries (i.e. horizontal) ^[1]. After three instructional meetings to align disagreements on a trial set of cases, the two graduate assistants proceeded autonomously in the classification activity and discussed any differences with the authors to reach a consensus. To increase the transparency and replicability of such a classification procedure, we provide in Appendix B a table showing some illustrative examples of how we assigned a strategic focus (vertical or horizontal) to the corporate accelerator in our dataset.

Table X presents a selection of four corporate accelerators, specifying their corporate sponsor, industry, and textual evidence (from corporate websites and public databases such as Crunchbase) used to support the classification. This approach allowed us to systematically analyze each accelerator’s mission statements, focus areas, and eligibility criteria to infer whether the accelerator’s strategic scope was narrow and aligned with the sponsor’s domain (vertical) or broad and cross-sectoral (horizontal). For instance, The Disney Accelerator, sponsored by The Walt Disney Company, emphasizes its goal of supporting startups that contribute to the future of technology and entertainment—clearly aligned with the company’s core business—therefore classified as vertical. Similarly, the Merck Accelerator, run by Merck and focused on healthcare, life sciences, and performance materials, prioritizes startups operating in fields directly linked to its innovation domains, such as biosensing, AI-enabled health, and clean meat. This specific alignment supports its classification as vertical.

Conversely, Google for Startups Accelerator includes multiple thematic programs (e.g. AI for energy, Canadian tech startups) that span industries well beyond Google’s advertising and Internet services core. Its open, cross-industry approach justifies the horizontal classification. Finally, Chivas Venture, sponsored by Chivas Brothers, supports social entrepreneurs from any sector working toward social and environmental impact. This focus on mission over market sector leads to its classification as horizontal.

To assess the performance of corporate accelerators we used two distinct dependent variables: Exit Rate and Survival Rate. Exit Rate measures the rate of startups that, within three years of completing the corporate acceleration program, have either undergone an initial public offering (IPO) or been acquired, relative to the total number of startups that participated in that program. As regards Survival Rate, it measures the ratio between the number of startups still in business after three years from the end of the corporate’s acceleration program and the total number of startups accelerated by that program. Our choice to use a three-year observation window is aligned with the most commonly adopted standard in the literature on startup performance and accelerator impact (e.g. Hallen et al., 2023; Yu, 2020; Seitz et al., 2024; Avnimelech et al., 2025; Del Sarto et al., 2022; Fehder, 2024; Assenova and Raphael, 2024).

Considering the independent variables, we measured the Management structure of the corporate accelerators through a binary variable, which assumes a value of 1 if the corporate accelerator is in-house, and 0 if it is a powered-by corporate accelerator (Moschner et al., 2019). Moreover, following previous literature (Seitz et al., 2024), we assessed the Strategic focus of the accelerators using a binary variable that assumes value 1 if the acceleration program is exclusively dedicated to startups operating in industries related to the core business of the corporation (i.e. vertical), 0 otherwise (i.e. horizontal).

Furthermore, we included in our model several control variables. We controlled for the location of the Corporate accelerators’ headquarter by including three dummy variables, “USA”, “Europe”, and “Rest of the world”, which assume value 1 if the headquarter of the corporate accelerator is located in that geographical area, 0 otherwise. Furthermore, we added a dummy control variable, Ecosystem maturity, to capture the maturity of the entrepreneurial ecosystem where each corporate accelerator is located. Specifically, following established methodology from previous studies we identified the more mature startup ecosystems worldwide using the 2023 Global Startup Ecosystem Ranking ^[2] (Berger and Kuckertz, 2016), and operationalized the control variable Ecosystem maturity as a binary variable that assigns a value of 1 if the accelerator is situated in one of the top 10 mature ecosystems (i.e. Silicon Valley, New York City, London, Los Angeles, Tel Aviv, Boston, Beijing, Singapore, Shanghai, and Seattle), and 0 otherwise. We also controlled for the Age of the accelerator including a variable measuring the natural logarithm of the number of years passed since the corporate accelerator foundation. Moreover, we controlled for the length of the acceleration program; the variable Duration measures the natural logarithm of the number of weeks of the acceleration program. In the two aforementioned cases, the logarithmic transformation has been applied to address issues related to the skewness of the variables’ distribution (Bartlett, 1947). In addition, we controlled for the Size of the corporate backing the accelerator through a categorical variable as declared in Crunchbase; Size assumes value 1 if the revenue of the is lower than $1M, 2 if it ranges between $1M and $10M, 3 if it ranges between $10M and $50M, 4 if it ranges between $50M and $100M, 5 if it ranges between $100M and $500M, 6 if it ranges between $500M and $1B, 7 if it ranges between $1B and $10B, and 8 if it is higher than $10B. Finally, we included a dummy variable to control for the stage of development of the startups accelerated (Hallen et al., 2023); Stage of development is a dummy variable assuming value 1 whether the corporate accelerator focuses on early-stage startups, 0 whether the corporate accelerator targets late-stage startups.

Table 1 reports descriptive statistics. The sample includes corporate accelerators primarily based in the USA (33.1%), Europe (37.8%), and other regions (29.1%). On average, they are 7 years old, run 15-week programs, and 43.9% target early-stage startups. The average revenue falls between $1M and $10M. Most accelerators (78.0%) are managed in-house, and nearly half (49.7%) adopt a vertical, exploitative focus. As regards the performance measures, the corporate accelerators in our sample show a mean exit rate equal to 0.130 and a survival rate for their startups equal to 0.884.

Table 2 details performance measures along corporate accelerators’ locations and main dimensions (i.e. Strategic focus and Management structure). Particularly, European accelerators report the highest average exit rates, while USA accelerators lead in survival rates. As regards the closure rate, accelerators headquartered in other countries than USA and Europe show the lowest levels. Programs in mature ecosystems outperform those in less developed ones on both metrics. In-house accelerators show slightly higher exit and survival rates and lower closure rates than powered-by models, and vertical accelerators slightly outperform horizontal ones.

Table 3 outlines the characteristics of corporate accelerators across different strategic focuses and management structures (as well as their interactions), examining their distribution by region, ecosystem maturity, age, duration, and size. In-house accelerators are more prevalent in mature ecosystems and tend to be older, larger, and have longer program durations compared to powered-by ones. The sample includes more Europe-based accelerators, both in-house and powered-by. Regarding strategic focus, horizontal accelerators are more common in mature ecosystems, with greater average age, size, and duration than vertical ones. The sample also shows a higher share of vertical accelerators in Europe, while horizontal accelerators are more frequently based outside the USA and Europe.

Figure 1 presents four pie charts illustrating the distribution of 6,444 startups accelerated by 188 corporate accelerators, categorized by region, ecosystem maturity, management structure, and strategic focus. Most startups were accelerated by U.S.-based programs, with smaller shares in Europe and the rest of the world. The majority were located in non-mature ecosystems and participated in in-house accelerators. Finally, a larger portion of startups was accelerated by vertical rather than horizontal programs.

Table 4 reports the pairwise correlations, which show no critical issues. Additionally, VIF values confirm the absence of multicollinearity concerns, indicating that all variables can be included in the models simultaneously (Stevens, 2012; Gujarati, 2004).

Since our dependent variables are continuous, we employed Ordinary Least Squares (OLS) regression model. Results are shown in Table 5 for Exit Rate, and Table 6 for Survival Rate. Using a hierarchical approach, we assessed the incremental impact of each independent variable. Particularly, Models 1 and 6 include only control variables and serve as baseline models; Models 2 and 7 add Management Structure; Models 3 and 8 introduce Strategic Focus. Models 4 and 9 include both variables. Finally, Models 5 and 10 test their combined effect by adding the interaction term Strategic Focus^*Management Structure. This stepwise approach ensures robustness and helps isolate the contribution of each variable.

Considering Table 5 and starting with the control variables, we focus on Model 1. The variable Duration is significant and has a positive coefficient meaning that the Exit Rate is positively affected by the length of the acceleration program. Moreover, the control variable Age is significant and has a positive coefficient suggesting that more experienced accelerators are able to reach higher values of Exit Rate. Considering the dummy variables related to the location of accelerators headquarter, “USA” is not significant while “Europe” exhibits a positive and significant coefficient. This result suggests that corporate accelerators located in “Europe” are characterized by higher values of Exit Rate that those in the “Rest of the word” (omitted as baseline category). The control variable related to the investment stage of accelerators, Stage of development, is significant and negatively affects the number of Startup exits meaning that accelerator programs that support startups in their early stage of development achieve lower performance compared to those focusing on late-stage startups. Finally, Ecosystem maturity and Size are not significant in explaining the Exit Rate of corporate accelerators. Moving on to the independent variables, in Model 2, Management structure is significant and has a positive impact on the Exit Rate, supporting the idea that in-house accelerators (internal management structure) reach a better performance than the powered-by ones. Considering Model 3, the dependent variable Strategic focus is significant and has a negative coefficient, suggesting that the decision to take a vertical approach hurts the corporate accelerators’ Exit Rate. Model 4 further corroborates the key relationships between the two independent variables and the dependent variable Exit Rate identified in the previous models. Finally, in Model 5 the interaction term Management Structure*Strategic focus is significant and has a negative coefficient suggesting that the combined effect of an internal management structure and a close fit with the corporations’ core business may reduce the exit rate of startups. This implies that the presence of both characteristics (i.e. in-house vertical corporate accelerators) limits the corporate accelerators’ Exit Rate.

Referring to Table 6 and beginning with the control variables in Model 6, we observe that the variables Duration, Ecosystem maturity, and Size are not significant. The control variable Age is significant with a positive coefficient, suggesting that more experienced accelerators achieve higher level of Survival rate for their startups. The dummy variables related to the accelerator’s headquarters location, “USA” and “Europe” are not significant. The control variable Stage of development, is significant and positively impacts the corporate accelerators’ Survival rate, meaning that programs supporting early-stage startups exhibit higher performance than those focused on late-stage startups. Turning to the independent variables, in Model 7, Management structure is significant and positively affects Survival rate, supporting the idea that in-house accelerators (internal management structure) perform better than those powered-by. In Model 8, the variable Strategic focus is significant with a positive coefficient, suggesting that a vertical focus increases the Survival rate of corporate accelerators. Model 9 reinforces the evidence supporting the main relationships established in the earlier models between the two independent variables and the dependent variable Survival rate. Finally, in Model 10, the interaction term Management Structure * Strategic Focus is significant with a negative coefficient, indicating that the combination of an internal management structure and alignment with the corporation’s core business may reduce startup survival. This suggests that in-house vertically focused corporate accelerators are associated with a lower Survival rate.

Additional analyses were carried out to check the robustness of previous results. Particularly, we assessed the hypotheses by using two alternative dependent variables, Adjusted survival rate and Closure rate.

While Survival Rate includes both startups that remain active and those that have exited, Adjusted Survival Rate excludes exited firms and considers only those that are still active at the end of the observation window. This measure allows us to isolate firms that continue to operate independently, thereby offering a more conservative view of startup survival. The variable is calculated as the number of startups in each accelerator that remained active three years after program completion, relative to the total number of startups initially accelerated by the accelerator.

The dependent variable Closure Rate captures the proportion of startups that ceased operations within the observation window. This variable reflects the share of startups that were definitively closed three years after completing the acceleration program, providing an unambiguous indicator of negative performance outcomes, since closure represents a definitive end to the startup’s activity. Closure Rate is calculated as the ratio between closed startups and the total number of startups accelerated by each accelerator.

The results of these additional analyses, which are reported in Tables VII and VIII (Appendix A), are consistent with those obtained with the variable Survival rate (Table 6).

Moreover, to further examine the combined effect of structural and strategic focus, we created four binary variables reflecting the combination of accelerator management structure and strategic focus: In-house Vertical, In-house Horizontal, Powered-by Vertical, and Powered-by Horizontal. Then, using these variables, we performed a regression model, using In-house Horizontal as the reference category. The results, shown in Table IX (Appendix A), reinforce those obtained in the main analysis (Tables 5 and 6) confirming that In-house Horizontal accelerators are consistently associated with higher startup exit and survival rates compared to In-house Vertical programs. As such the comparison across accelerator configurations provides further support for the explanatory power of the interaction effect between the two accelerators’ dimension under investigation as observed in our main models.

Our findings indicate that variations in the corporate accelerators’ performance can be attributed to their management structures and strategic focuses. Particularly, our findings suggest that in-house corporate accelerators provide stronger support for startup exit and survival than those powered by external entities. This may reflect the benefits of close collaboration between corporations and startups that in-house models enable, compared to the broader network access and investor exposure offered by powered-by accelerators. In-house accelerators, often located near the parent company’s headquarters, allow startups to leverage corporate resources and foster frequent, informal interactions that promote knowledge exchange, trust, and ultimately, survival and growth (Moschner et al., 2019). Additionally, they may serve as learning investments for corporations, aligning with a Real Options perspective (Folta, 1998; Folta and Miller, 2002), by allowing firms to monitor startup development and adjust their commitment, such as through acquisition, once uncertainty about the startup’s technology diminishes.

Moreover, our findings show that the corporate accelerator’s strategic focus significantly influences performance outcomes, albeit in different ways. Specifically, vertical accelerators are more strongly associated with improved startup survival, whereas horizontal accelerators are linked to higher exit rates. These results do not indicate that one model is inherently superior to the other. Rather, they suggest that vertical and horizontal accelerators may be optimized for different strategic goals. Vertical accelerators, where startups and corporate sponsors operate in closely related technological or industrial domains, can facilitate deeper knowledge transfer and sustained support. This alignment enhances the corporation’s ability to evaluate startups’ potential and provide targeted, domain-specific resources, thereby increasing survival prospects (Kohler, 2016; Moschner et al., 2019; Ivanov and Xie, 2010). Furthermore, technological proximity strengthens startups’ absorptive capacity, enabling them to better integrate corporate knowledge and capabilities (Cohen and Levinthal, 1990; Mowery et al., 1996). On the other hand, horizontal accelerators expose startups to a more diverse set of knowledge domains, potentially enhancing opportunities for exploration, serendipitous learning, and faster access to markets. This broader scope may make horizontal programs better suited to fostering interdisciplinary collaboration and resource variety, a condition conducive to exit (Noviaristanti et al., 2024).

Finally, our results underscore the critical importance of examining not only the individual effects of corporate accelerator design features but also their interaction. The inclusion of the interaction term between management structure and strategic focus leads to a substantial increase in the model’s explanatory power. This suggests that the interplay between these dimensions is a key determinant of startup performance, potentially more influential than each dimension considered in isolation. Specifically, we found that in-house corporate accelerators adopting an explorative strategic focus (i.e. horizontal) exhibit significantly higher exit and survival rates for their startups compared to in-house programs with an exploitative focus (i.e. vertical). We interpret this finding as evidence that an explorative strategic focus can mitigate structural limitations typically associated with in-house models. A well-known drawback of in-house accelerators lies in their limited exposure to diverse external knowledge, which may cause a lock-in effect (Mahmoud-Jouini et al., 2018). However, when these accelerators adopt a horizontal approach, startups gain access to a more heterogeneous set of knowledge domains, facilitating broader learning and innovation opportunities. The tight integration with the parent corporation and exposure to cross-industry knowledge can foster creative recombination and dynamic capabilities that ultimately enhance startup performance. Thus, the strong contribution of the interaction term reveals the synergistic value of aligning governance and strategic orientation in corporate accelerator design and highlights how such alignment can be a critical source of performance differentiation.

This paper offers several important theoretical insights. First, this study adds to the limited and recent research on the performance of corporate accelerators (Canovas-Saiz et al., 2020, 2021). While numerous studies have addressed the performance of accelerated startups (Gonzalez-Uribe and Leatherbee, 2018; Hallen et al., 2023; Yagüe-Perales et al., 2024; Yu, 2020), there remains a notable scarcity of research examining accelerators themselves as the primary unit of analysis. These studies have focused on some dimensions characterizing corporate accelerators including the age, geographical location, and portfolio size as key drivers of their performance (Canovas-Saiz et al., 2020, 2021). Thus, our paper extends this stream of literature by suggesting that, alongside these attributes, the management structure and the strategic focus can predict the success of corporate accelerators.

Moreover, previous research examining the relationship between the attributes of corporate accelerators and their performance has exclusively scrutinized individual dimensions in isolation, thus providing limited insight into the intricate interplay between various facets of these innovative corporate venturing modes. Our study adopts a different approach by concurrently investigating the simultaneous impact of corporate accelerators’ management structure and strategic focus. Such an approach allowed us to unveil the potential for a counterbalancing effect between different dimensions. Specifically, while the internal management structure may present inherent limitations for the success of corporate accelerators, our findings demonstrate that these shortcomings could be offset by the positive influences emanating from an explorative strategic focus. Through this result, our study challenges the current wisdom that views dimensions of corporate accelerators in silos and contributes to previous literature underscoring the importance of adopting a holistic perspective when exploring the corporate accelerators’ landscape.

Finally, this study adds to the existing literature investigating the effect of corporate venturing modes on startups’ exiting (Ivanov and Xie, 2010; Kim and Haemin, 2017; Wang et al., 2022). This research has focused on the impact of CVC investments on startup exits, elucidating the role of various contingency factors as timing and strategic alignment. For example, some researchers have shown that CVC investments can signal credibility, increasing a startup’s likelihood of exit, particularly when startups share knowledge and technology with their investing corporations, which enhances the chances of exits via IPOs or acquisitions (Maula and Murray, 2002; Ivanov and Xie, 2010). Conversely, other scholars have revealed that startups receiving investments during the early stages exhibit a lower propensity for IPOs (Kim and Haemin, 2017). Additionally, studies have indicated that startups receiving joint funding from a syndicate of VC investors are more likely to be acquired (Wang et al., 2022). Our results suggest that accelerator programs have varied impacts on exit strategies. As CVC investments, close relationships between corporations in in-house accelerators can enhance a startup’s exit prospects by signaling credibility. However, unlike CVC funding from corporates with matching knowledge and technology, strong alignment between startups and corporate partners in in-house accelerators may reduce exit chances. This study contributes to the understanding of startup exit strategies by integrating insights from corporate venturing literature with a detailed exploration of accelerator program dimensions.

Our research offers key recommendations for entrepreneurs assessing corporate accelerator programs, urging a strategic selection aligned with their goals. Choosing an accelerator goes beyond industry alignment, it’s about leveraging an ecosystem that supports survival or exit.

Entrepreneurs should assess both the management structure and strategic focus of accelerators, as these shape access to resources, networks, and long-term success. Particularly, for those aiming at exit, in-house corporate accelerators can be highly beneficial. Closely integrated with the parent firm, these programs offer privileged access to funding, expertise, and powerful networks. Startups leveraging these assets may accelerate their growth, making them attractive to acquirers.

However, entrepreneurs should also consider whether the accelerator takes an exploitative (aligned with corporate priorities) or explorative (fostering broader innovation) approach. Vertical exploitative models can foster dependency on the corporate sponsor, potentially leading to acquisition, but also reducing flexibility. In contrast, horizontal explorative programs offer startups more independence, benefiting those seeking long-term growth, cross-industry synergies, and adaptability.

Therefore, entrepreneurs must weigh trade-offs carefully considering these two accelerators’ attributes. While vertical programs may offer strong alignment and support, they can also limit autonomy and appeal to external investors. Selecting an in-house accelerator with an explorative approach allows startups to benefit from corporate resources while maintaining strategic flexibility and pursuing tailored growth or exit paths.

Lastly, our study advises corporate managers on program design. Acceleration efforts should align with corporate goals while remaining attractive to startups. This involves choosing the right management structure and focus to support startup success and potential exits. Demonstrating a strong track record and regularly sharing performance outcomes can enhance the accelerator’s appeal, ultimately advancing broader corporate objectives.

This study outlines some limitations that offer prospects for future investigations. The first limitation pertains to the performance measure used to assess the corporate accelerators’ dimensions and performance. While the chosen measures (i.e. the exit and survival rate) offer valuable insights into the understanding of accelerators’ success, it is crucial to acknowledge that alternative performance measures might yield different results. Indeed, different performance measures, such as the relationship-building in the post-acceleration phase (Kramer and Kanbach, 2023), the evaluation attained (Cohen et al., 2019), the capital raised by the startups (Avnimelech et al., 2025; Canovas-Saiz et al., 2021; Cohen et al., 2019; Yu, 2020), the status of the investors attracted (Hallen et al., 2023), and the revenue growth can capture distinct aspects of the phenomena under investigation. As such, future research endeavors should consider employing alternative performance measures to validate and complement the findings of this study, uncovering nuances and insights that might have been overlooked when relying on a single performance measure.

Moreover, we acknowledge that the management structure and the strategic focus measurements should benefit from additional granularity, which could enrich the analysis. In our study, we used a dichotomic categorization of corporate accelerators’ strategic focus as vertical or horizontal, and applied a binary classification to their management structures, aligning it with literature in the field that distinguishes between internal and external (i.e. in-house vs. powered-by) (Seitz et al., 2024; Moschner et al., 2019). This approach allows for consistency and comparability with existing studies and provides a clear basis for our analysis. However, future studies could expand upon this by investigating a broader spectrum within these classifications. For instance, future research could further refine the categorization of corporate accelerators by distinguishing between “horizontal-generalist” accelerators, which engage broadly across industries, and “horizontal-aligned” accelerators that focus on sectors more closely related to the parent corporation’s core business. This distinction may be important as it may influence the dynamics of collaboration and the potential for conflicts of interest between corporations and startups, ultimately influencing their performance. Particularly, while horizontal-aligned accelerators may offer greater value alignment, they may also present higher risks of conflicts over intellectual property or strategic objectives. Therefore, taking into account this fine-grained distinction, future studies can provide more nuanced insights into how the accelerators’ strategic focus influences their outcomes.

A further limitation of the study is the lack of data on startups before they participate in corporate accelerators. For instance, we do not have information on their initial funding levels, prior participation in other acceleration programs, or other indicators of ex ante startup quality. As a result, we cannot fully disentangle whether differences in performance stem from the value added by corporate accelerators or from the pre-existing quality of the startups they select or attract. Future research could address this limitation by integrating startup-level data before acceleration, allowing for a more precise assessment of the causal relationship between accelerator characteristics and startup outcomes.

Finally, another limitation of this study is the lack of large-scale, comprehensive datasets, which restricts the ability to conduct robust longitudinal analyses and limits the depth of insights into the performance of corporate accelerators over time. As a result, the findings may be constrained by the available data, and caution is needed in generalizing the results. Future research could address these limitations by conducting qualitative studies to gather additional insights to better understand accelerator success. For example, future studies can complement the empirical analysis by conducting qualitative interviews with a diverse range of participants involved in corporate accelerators (e.g. startup founders and accelerators’ managers), who can yield valuable insights into the influence of corporate accelerators’ attributes on their performance.

The supplementary material for this article can be found online.