We are sincerely grateful to Dr. Nguyen for the constructive and technically rigorous nature of her review. Her observations have been particularly valuable in strengthening the replicability, methodological transparency, and structural coherence of the manuscript. We have addressed each of her points in detail, and we believe the revised version now resolves the concerns she raised. Our point-by-point response follows.

Research background and objective

Comment. "You stated that this research aims to analyze existing metrics of entrepreneurial ecosystem performance. Still, you then said to include only general-purpose entrepreneurial ecosystems and exclude digital and university entrepreneurial ecosystems because of the different metrics used. With this in mind, the research objective and title should clearly indicate the type of entrepreneurial ecosystem intended to be reviewed to avoid misinterpretation."

Response. We thank the reviewer for this important observation regarding the alignment between the stated scope and the actual coverage of the review. We agree that the original formulation generated an interpretive ambiguity that needed to be resolved. Three modifications have been introduced in response. First, the title has been refined to specify the scope of the review and prevent misinterpretation. Second, the research objective stated in the Introduction now explicitly delimits the analysis to general-purpose, territorially anchored entrepreneurial ecosystems, distinguishing them from sectoral, digital, and university-based variants. Third, the Methods section provides an analytically grounded justification for this delimitation—rooted in the categorically distinct measurement logics that govern digital ecosystems (platform-mediated dynamics, network externalities, algorithmic governance) and university ecosystems (institutional missions, knowledge-transfer mandates, academic-entrepreneurship metrics). The reader is therefore informed of the precise scope of the review from the outset, both in the title and in the framing of the objective.

Methods — Justification of databases, complementary terms, and tools

Comment. "You justified the significance of the Scopus and Web of Science databases, as well as the complementary terms used, without citing any sources […]. The introduction of Rayyan software, VOSViewer, and PRISMA, as applied in the research, also lacked sources, i.e., what they are and why they were considered for use."

Response. This observation has been addressed in full. Every methodological justification in the revised Methods section is now supported by appropriate scholarly sources. Specifically:

The selection of Scopus and Web of Science as the primary databases is now substantiated with reference to the bibliometric literature on database coverage, indexing standards, and complementarity for systematic reviews (Pranckutė, 2021; Torres-Salinas & Arroyo-Machado, 2026; Ciule et al., 2025).

Methods — Bias mitigation and reliability of inclusion and analysis

Comment. "You did not justify how you mitigated bias or ensured the reliability of including the articles for review and of the data analysis. This information is essential for a systematic review as it tells the readers how reliable your review is."

Response. We accept this critique and have introduced a dedicated subsection on quality assurance, bias mitigation, and inter-rater reliability in the revised Methods section. The procedures now explicitly documented include: (i) blinded dual screening of titles and abstracts in Rayyan, with both reviewers operating independently and disagreements flagged automatically by the platform; (ii) consensus resolution of discrepancies through structured discussion, with a third reviewer adjudicating unresolved cases; (iii) inter-rater reliability assessment using Cohen's κ on the screening decisions, with the obtained value reported in the manuscript; (iv) full-text dual review of the records advancing to the eligibility stage, applying the same dual–independent procedure; and (v) explicit documentation of exclusion reasons at each stage, fully traceable through the PRISMA flow diagram (see next point). For the data-analysis stage, we have added a description of the cross-validation procedures applied to the thematic coding of metrics and to the interpretation of clusters, including triangulation between the two coders and reconciliation through structured comparison against the scientific-background table now incorporated in the manuscript.

Results and Discussion — PRISMA flow diagram

Comment. "The PRISMA flow diagram (Figure 1) should be moved to the methods section and should show clearly the reasons for excluding the articles including the number of articles, for example, foreign language, research topic, etc. You should also use the standard PRISMA flow diagram, available at [the official PRISMA 2020 link]."

Response. We thank the reviewer for this technically precise correction. The PRISMA flow diagram has been (i) reproduced using the standard PRISMA 2020 template available at the official PRISMA Statement website, ensuring full conformity with the current reporting guidelines, (ii) relocated to the Methods section, where it now belongs in accordance with PRISMA 2020 recommendations, and (iii) supplemented with disaggregated exclusion reasons at each filtering stage—including, among others, language restrictions, topical mismatch (digital and university ecosystems), document type (literature reviews, editorials, conference abstracts), and absence of full-text availability—with the corresponding numerical breakdown for each category. The revised diagram therefore provides a transparent, fully auditable record of the screening process.

Results and Discussion — Structural coherence around metrics

Comment. "The results and discussion are really problematic and rambling. The discussion should present research findings, but in your research, it did not reflect the results; it even introduced new ideas. The main focus of your study is the metrics; therefore, the results and discussion should be structured around them. What was the connection between the clusters and the metrics mentioned in Table 3? Also, what was the connection between the methodologies and the metrics? You provided many pieces of information, but they looked scattered and did not address what the research was actually seeking to answer."

Response. We accept this observation in full and have undertaken a substantial structural reorganisation of the Results and Discussion sections to ensure that metrics function as the consistent analytical axis throughout. The revised structure now operates as follows:

The Results section presents the bibliometric and thematic findings explicitly tied to metrics: each thematic cluster identified through VOSviewer is now linked to the families of metrics dominant within it (e.g., the regional-development cluster is connected to fsQCA-based and territorial-policy indicators; the global-measurement cluster to composite indices such as GEI and REDI; the network-dynamics cluster to relational and brokerage metrics). The reader can therefore trace, cluster by cluster, the analytical journey from bibliometric structure to specific measurement instruments.

Methods — Replicability of the VOSviewer co-occurrence analysis

Comment. "There needs to be elaboration on the use of VOSViewer to generate a co-occurrence map to increase the replicability of your research. Did you use a keyword thesaurus? How did you set the thresholds? Were the clusters generated automatically, or were they generated manually?"

Response. We are grateful for these technically precise questions, which directly address the replicability of the analysis. The revised Methods section now provides full disclosure of the VOSviewer procedure, addressing each of the reviewer's specific queries:

Unit of analysis: author keywords combined with indexed keywords (KeyWords Plus and Index Keywords from WoS and Scopus, respectively).

Conclusion — Future research directions

Comment. "There needs to be a mention of future research directions or an agenda based on the results and research gaps identified; otherwise, a review looks like a summary of what has been researched."

Response. We fully agree. The Conclusion has been substantially expanded to articulate a concrete and actionable research agenda derived from the gaps identified in the review. Specifically, four research pathways are now outlined: (i) the development of recursive panel designs capable of capturing the temporal dynamics and condition–outcome cycles intrinsic to EEs; (ii) the systematic deployment of network-analytic operationalisations using founder-investor-mentor data from sources such as Crunchbase and PitchBook; (iii) the application of configurational approaches (fsQCA) to identify equifinal pathways through which heterogeneous ecosystems achieve comparable performance; and (iv) the incorporation of digital-trace metrics (online attention, platform visibility, attention-flow data) as a high-frequency complement to traditional indicators. Each pathway is accompanied by specific recommendations regarding data sources, analytical techniques, and identification challenges. The Conclusion thus moves beyond summary toward a forward-looking research agenda anchored in the gaps surfaced by the review itself.

Paper structure — Paragraph consolidation

Comment. "In addition to the structural problems mentioned, you had too many small paragraphs in your manuscript, which scattered the information and reduced the flow and consistency of the research. I recommend you restructure your manuscript, at least by linking the ideas into longer paragraphs."

Response. We thank the reviewer for this observation, which was also raised by Reviewer 1 and which we have taken very seriously. The entire manuscript has been restructured: short, isolated paragraphs have been consolidated into cohesive analytical units that integrate the related ideas, supporting evidence, and citations into a continuous line of reasoning. This restructuring is particularly visible in the Introduction, the conceptual section, the Discussion, and the Conclusion, where the consolidated paragraphs now provide a more substantive, fluent, and academically appropriate exposition.

We are deeply grateful to Dr. Nguyen for the precision and constructiveness of her review. Her observations have measurably improved the replicability, transparency, and structural coherence of the manuscript, and we trust that the revised version now meets the standards she rightly demands of a rigorous and replicable systematic review.

We are sincerely grateful to Professor Ahmed for the depth, rigour, and intellectual seriousness of his evaluation. His critical reading has compelled us to reconceptualise several aspects of the manuscript and has materially elevated its analytical ambition. Below we address each point in detail.

Comment. "The manuscript remains largely superficial in its analysis and contribution. The study primarily catalogs existing metrics without offering substantive theoretical integration, methodological advancement, or novel analytical insight […] compliance with protocol alone does not compensate for the lack of analytical depth."

Response. We accept the substance of this critique. The previous version privileged procedural transparency over interpretive synthesis, and the cataloguing of metrics was not adequately translated into an integrative analytical framework. The revised manuscript has been substantially reworked to move beyond description toward synthesis through three structural changes: (i) a new conceptual section, located before the Methods, debates the EE construct and consolidates its evolution into a synoptic table articulating measurement implications across four developmental phases; (ii) the classification of metrics is now interpreted through a critical lens that problematises causal ambiguity, temporal sequencing, and the dual condition–outcome status of certain indicators; and (iii) the Discussion has been rewritten to articulate the contribution of the review as an integrative reading of how measurement choices reflect and reproduce underlying theoretical commitments.

Comment. "The introduction reiterates familiar arguments regarding the importance of EEs and the lack of consensus on their measurement […] The manuscript does not clearly articulate how it advances beyond existing reviews, nor does it identify a precise conceptual or theoretical gap that this study resolves."

Response. This observation has prompted a thorough revision of the conceptual framing. The Introduction has been rewritten to identify a precise and previously under-addressed gap: although prior reviews have catalogued metrics or mapped the conceptual evolution of EEs, none has systematically interrogated the epistemological coherence between the theoretical conceptualisation of EEs (as relational, dynamic, systemic constructs) and the measurement instruments deployed to assess them (which remain predominantly static, aggregative, and input-output oriented). Our revised contribution is reframed not as an additional catalogue but as a critical synthesis exposing this theory–measurement misalignment and a structured roadmap for resolving it. The revised Introduction explicitly differentiates our contribution from existing reviews (e.g., Cavallo et al., 2019; Theodoraki et al., 2024).

Comment. "The exclusion of digital and university-based ecosystems is asserted rather than analytically justified […] the authors do not engage with the conceptual implications of this decision or reflect on how it limits the scope and interpretability of their findings."

Response. We accept this observation in full. The revised version dedicates a paragraph in the Methods, with parallel reflection in the Discussion, to justify the exclusion on conceptual grounds: digital ecosystems are governed by platform-mediated logics generating measurement requirements (e.g., platform stickiness, multi-sided market indicators, digital trace data) categorically distinct from those applicable to territorially anchored, general-purpose EEs (Bejjani et al., 2023; Pigola et al., 2024); university-based ecosystems, in turn, operate under institutional missions and academic-entrepreneurship metrics (e.g., spin-off rates, patent licensing) responding to a distinct theoretical tradition (Ayala-Gaytán et al., 2024). The Limitations section acknowledges that this choice circumscribes interpretability to general-purpose territorial EEs and invites future reviews to undertake parallel syntheses for adjacent constructs.

Comment. "The methodological section is procedurally sound and transparently reported […] However, methodological correctness should not be conflated with methodological rigor in analysis."

Response. We thank the reviewer for this important distinction, which we now explicitly incorporate. The Methods section has been complemented by a new subsection on analytical strategy, articulating how we move from procedural compliance to interpretive synthesis through (i) thematic re-coding of clusters against the theoretical phases identified in our scientific-background table, (ii) critical mapping of measurement instruments against the conceptual claims they purport to operationalise, and (iii) explicit attention to the theoretical and methodological tensions underlying the metrics inventoried.

Comment. "The search strategy is narrow and conservative, relying heavily on title-based keyword filtering. This design choice likely excludes a substantial body of relevant work that addresses ecosystem performance using adjacent or alternative terminology."

Response. We respectfully clarify that the title-restricted filtering on the EE construct was deliberately combined with a broader title–abstract–keyword search on performance terms (index OR performance OR metric OR indicator**), which substantially mitigates the restrictiveness suggested. The decision to anchor the EE term in titles ensured topical centrality and avoided the noise generated by tangential mentions in abstracts. That said, the reviewer is correct that this constitutes a trade-off deserving explicit treatment. The revised Methods section now (i) discloses this trade-off transparently, (ii) reports the results of a sensitivity analysis extending the search to title–abstract–keyword filtering, and (iii) frames this design choice in the Limitations section as a deliberate prioritisation of precision over recall, consistent with the analytical orientation of the review.

Comment. "Bibliometric outputs such as publication trends and keyword co-occurrence maps are presented, but they are not analytically exploited […] the cluster analysis remains a mapping exercise rather than an interpretive one."

Response. We accept this critique. The Results section has been substantially rewritten to render the cluster analysis genuinely interpretive: (i) each cluster is mapped onto the developmental phases in our scientific-background table, exposing the theoretical lineage of its dominant metrics; (ii) inter-cluster relationships are explicitly examined—identifying, for instance, the tension between the quantitative-benchmarking cluster (composite indices) and the relational-process cluster (network analyses), which operationalise incommensurable conceptions of ecosystem performance; and (iii) the underlying theoretical assumptions of each cluster are surfaced and critically discussed. The bibliometric analysis is thus repositioned as a structuring device for theoretical interpretation rather than a stand-alone descriptive output.

Comment. "The classification of metrics into inputs, outputs, composite indices, subjective measures, and emerging indicators is intuitive but analytically shallow. Similar classifications already exist in the literature, and the manuscript does not convincingly demonstrate how its framework differs from or improves upon them."

Response. The revised manuscript explicitly situates our classification within the prior typological landscape (e.g., Stam, 2015; Leendertse et al., 2022; Nicotra et al., 2018), articulating the specific points of departure. Our classification is distinguished by an additional analytical dimension—an epistemological reading of each category that identifies (i) the underlying theoretical tradition it operationalises, (ii) the temporal logic it embeds (static, lagged, dynamic), and (iii) its functional role within the EE causal structure (condition, mechanism, outcome). This dimension transforms the typology from a catalogue into an analytical instrument.

Comment. "The paper fails to grapple with core conceptual issues such as causal ambiguity, temporal sequencing, and endogeneity among ecosystem variables. Metrics that function simultaneously as conditions and outcomes are treated as discrete categories without sufficient theoretical justification."

Response. We regard this as the single most consequential critique in the review. The revised Discussion contains a dedicated section addressing these issues: we explicitly acknowledge the condition–outcome duality of several indicators—high-growth firms, for instance, function simultaneously as outputs of ecosystem performance and as inputs to its subsequent reproduction (Mukiza et al., 2020)—and discuss the implications of temporal sequencing and the endogeneity problems this generates. We argue that resolving these tensions requires longitudinal panel designs, instrumental-variable strategies, and dynamic measurement frameworks capable of modelling EEs as recursive systems in which today's outcomes become tomorrow's conditions.

Comment. "The discussion reiterates well-established critiques of ecosystem measurement, such as fragmentation and the dominance of static indicators. However, these observations are neither novel nor developed into a coherent analytical argument."

Response. The Discussion has been entirely rewritten around a unifying analytical thesis: that the persistent fragmentation of EE measurement is not a remediable defect of the literature but a symptom of a deeper theory–measurement misalignment, in which ecosystems are conceptualised relationally and dynamically yet measured aggregatively and statically. From this thesis we develop a coherent argument linking the fragmentation observed bibliometrically to the conceptual issues of causal ambiguity and temporal sequencing, and finally to the actionable research agenda articulated below.

Comment. "The proposed future research agenda is generic and lacks specificity. Calls for longitudinal studies, dynamic metrics, or mixed methods are commonplace in the field and do not provide actionable guidance."

Response. The revised agenda is organised around four concrete pathways, each with explicit methodological and data-strategy recommendations. First, recursive panel designs that operationalise EEs as time-indexed systems, drawing on longitudinal infrastructures (GEM, Eurostat regional panels, Crunchbase) and applying dynamic panel estimators (system-GMM, local projections). Second, network-analytic operationalisations using founder-investor-mentor co-occurrence data, with specific structural metrics (eigenvector centrality, brokerage, k-core decomposition) tied to theoretically motivated hypotheses. Third, configurational designs (fsQCA) to identify equifinal pathways through which heterogeneous ecosystems achieve comparable performance. Fourth, digital-trace metrics as high-frequency complements to traditional indicators, with explicit guidance on data sources and triangulation strategies.

Comment. "Despite procedural correctness, the manuscript lacks the analytical depth, conceptual originality, and critical synthesis required of a SLR. Addressing these shortcomings would require substantial reconceptualization and reanalysis, rather than incremental revision."

Response. We have undertaken precisely the substantial reconceptualisation and reanalysis the reviewer recommended. The revised manuscript reflects a meaningful elevation in conceptual ambition, analytical depth, and critical synthesis. We are deeply indebted to Professor Ahmed for the seriousness of his engagement and trust that the present version meets the scholarly standard he rightly demands.

We sincerely thank Professor Batista-Canino for her thoughtful and generous reading of our manuscript, as well as for her encouraging assessment of the topic and its relevance to the field. Her constructive observations have substantially strengthened the clarity, replicability, and analytical depth of the revised version. Below we provide a point-by-point response to each of her suggestions, indicating the corresponding modifications introduced in the manuscript.

Comment 1 — Paragraph structure: It would be desirable for the paragraphs to have greater substance, bringing together all the key information. Very short paragraphs (two or three lines) tend to reduce the quality of exposition and make the article feel somewhat telegraphic.

Response. We fully agree with the reviewer. The fragmented paragraph structure of the previous version compromised the argumentative flow and gave the manuscript a telegraphic tone. The entire article has therefore been carefully restructured: short, isolated paragraphs have been consolidated into cohesive analytical units that integrate the related ideas, supporting evidence, and citations into a continuous line of reasoning. This revision is particularly visible in the Introduction, the Discussion, and the Conclusion, where the consolidated paragraphs now provide a more substantive and fluent exposition. We are grateful to the reviewer for this observation, which has clearly improved the readability and scholarly tone of the work.

Comment 2 — Introduction of the concept: Since the literature review focuses on how entrepreneurial ecosystems are being measured, the article would benefit from starting with a discussion of the concept itself, rather than presenting it as an outcome of the review. I would suggest moving the section that debates the concept before the methodology and concluding it with a summary table of the concepts found in the literature.

Response. We thank the reviewer for this insightful suggestion, which has reshaped the conceptual architecture of the article. The discussion of the entrepreneurial ecosystem (EE) construct has been relocated to the front of the manuscript and now precedes the Methods section, providing the reader with the necessary theoretical grounding before the methodological procedure is introduced. The reorganised section opens with a critical debate on the conceptual evolution of EEs—from the foundational ecological metaphor through the systemic and relational turns—and culminates in a synoptic table that summarises the principal conceptualisations found in the literature, their key proponents, and their distinctive contributions. This modification clarifies the conceptual gap that the systematic review subsequently addresses and aligns the manuscript with the structural conventions of high-quality literature reviews.

Comment 3 — Methodological details: In the methodology section, the authors mention using VOSviewer, but it is important to specify the parameters applied to obtain those results. This is particularly useful to allow other researchers to replicate the process or extend the review in future publications, making the results comparable.

Response. We appreciate this important methodological observation, which directly addresses the replicability of the review. The Methods section has been substantially expanded to disclose the full set of parameters applied in VOSviewer (Van Eck & Waltman, 2010), including: (i) the unit of analysis (author keywords and indexed keywords), (ii) the counting method (full counting), (iii) the minimum co-occurrence threshold applied to the keywords retained for the network, (iv) the normalisation method (association strength), (v) the clustering algorithm and resolution parameter, and (vi) the procedures applied to consolidate semantic variants through a manually curated keyword thesaurus. Equivalent details have been added regarding the use of Rayyan (Ouzzani et al., 2016) for screening and the application of the PRISMA 2020 protocol (Page et al., 2021). These additions ensure that the entire pipeline can be independently reproduced and extended by future researchers.

Comment 4 — Scientific background: It would also be interesting to include a table summarizing the scientific background of the field to further strengthen the exposition. An example article is attached for reference.

Response. We are grateful for this suggestion and for the reference article provided, which served as a valuable model. A new synoptic table summarising the scientific background of the entrepreneurial ecosystem field has been incorporated into the revised manuscript. The table organises the intellectual evolution of the field into four developmental phases: (i) the conceptual foundations and ecological metaphor (Isenberg, 2010; Moore, 1993); (ii) the systemic shift and structural frameworks (Mason & Brown, 2014; Stam, 2015); (iii) indexation and quantitative benchmarking (Acs et al., 2017; Szerb et al., 2019); and (iv) the relational and process turn (Spigel, 2017; Spigel & Harrison, 2018). For each phase, the table identifies the key authors, their contribution to the field, and—most importantly for the purpose of this review—their specific implications for the measurement of EE performance. This addition strengthens the theoretical grounding of the manuscript and explicitly situates our review within the broader trajectory of the discipline.

Comment 5 — Future research avenues: In the discussion, it would be expected that the authors not only describe the results but also propose potential directions for future research. For instance: Which metrics seem most promising? Which could offer better results based on their knowledge and review?

Response. We thank the reviewer for highlighting this important shortcoming. The Discussion has been substantially expanded to articulate a concrete and actionable future research agenda derived from the evidence synthesised in the review. Specifically, we now identify the families of metrics that, in our judgement, hold the greatest analytical promise: (i) network-based and relational indicators, which appear best suited to capture the structural and dynamic nature of EEs; (ii) hybrid composite indices that integrate input, output, and perceptual dimensions while explicitly modelling temporal evolution; and (iii) emerging digital-trace metrics—such as online attention and platform-based visibility—that complement traditional indicators with high-frequency data. From these promising avenues, we derive specific methodological recommendations for future research, including longitudinal designs capable of disentangling causal sequencing, mixed-method strategies that combine fsQCA with bibliometric and network analyses, and cross-regional comparative studies aimed at testing the contextual transferability of measurement frameworks.

We are deeply grateful to Professor Batista-Canino for her constructive engagement with our work and for the favourable overall evaluation. Her recommendations have measurably improved the conceptual framing, methodological transparency, and scholarly contribution of the manuscript, and we trust that the revised version now meets the standards required for indexing.