Beyond Aggregate Success

Purpose. This study conducts a secondary distributional analysis of administrative gross merchandise value (GMV) records from the Agromarketing Masterclass TikTok Shop Edition (AMTTSE), a government-supported social-commerce programme in Malaysia. It examines whether aggregate sales were broadly shared across participating sellers.

Design/methodology/approach. A retrospective analysis of programme-administrative and seller-level GMV records for 80 companies and 160 entrepreneurs participating in AMTTSE from June to December 2024. The study applies descriptive distribution analysis, concentration measurement (top-k shares, Gini coefficient, Lorenz curve), seller segmentation and sensitivity analysis.

Findings. Cumulative sales reached RM6,205,957.06, but realised outcomes were highly concentrated: the top 10 sellers accounted for 52.4 per cent of December GMV and the seller-level Gini coefficient was approximately 0.71. Of 80 companies, 25 recorded zero GMV while 55 were active. Short-video (44.6 per cent) and livestream (35.4 per cent) channels dominated cumulative value capture. The paper treats monthly correlations as exploratory because of the small number of observations and the part-whole structure of the channel measures.

Originality/value. The article is materially distinct from two prior publications arising from the same programme: it addresses a new research problem (distributional value capture rather than programme effectiveness or capability relationships), uses a seller-level concentration lens, and adds Gini, Lorenz, segmentation and sensitivity analysis. Prior publications are disclosed transparently.

Keywords: value capture; social commerce; digital entrepreneurship; outcome inequality; government programmes

JEL classification: L26; M31; O33

Social commerce platforms have transformed the structure of market access for micro, small and medium-sized enterprises (MSMEs), combining entertainment, community interaction, product discovery and transaction functions within a single digital environment. For firms in emerging economies, this shift is consequential because traditional market access is often constrained by distribution cost, geographic distance, limited promotional capability and dependence on intermediaries. TikTok Shop has become particularly relevant because it allows sellers to combine short-video content, livestreaming, creator-based trust and platform-based fulfilment to reach consumers beyond conventional physical markets (Hajli, 2015; Wongkitrungrueng & Assarut, 2020).

Yet platform participation does not necessarily result in equal economic outcomes. Firms exposed to the same digital platform, the same training and the same institutional support may capture markedly different levels of commercial value. This heterogeneity is not an implementation detail; it is a central question for entrepreneurship and strategic management research. If the rationale for public investment in digital entrepreneurship is to extend market access to firms that would otherwise be excluded, then a programme whose benefits accrue overwhelmingly to a small group of sellers has not achieved broad-based inclusion even if its aggregate statistics are impressive (Lepak, Smith & Taylor, 2007; Gans & Ryall, 2017).

Why this problem matters theoretically is that it separates value creation from value capture. A firm may create substantial commercial activity without capturing commensurate economic value, and firms differ in their capacity to appropriate the value their activities create. The distinction between creating commercial activity and capturing economic value provides a more original theoretical framing than treating digitalisation as a single-stage process. The relevant question is not merely whether livestreaming or short video works, but why some firms can mobilise these formats more effectively than others (Teece, 2018; Eisenhardt & Martin, 2000).

Previous social-commerce research has predominantly examined consumer purchase intentions, trust, parasocial interaction and engagement (Sun et al., 2019; Wongkitrungrueng & Assarut, 2020; Lu & Chen, 2021). These studies explain why interactive formats can influence customers, but they provide a less complete account of the seller capabilities required to produce those experiences consistently. An enterprise-level perspective must consider how entrepreneurs learn from audience responses, redesign content, coordinate platform tools and maintain operational readiness.

What remains insufficiently understood is the distribution of commercial outcomes among firms participating in the same platform-enabled entrepreneurship intervention. Aggregate programme metrics conceal the extent to which value is concentrated among a small group of sellers. This gap is especially pronounced in government-supported entrepreneurship programmes, where evaluation tends to emphasise participant numbers, training completion and aggregate sales rather than the distribution of gains.

Government-supported entrepreneurship programmes create a theoretically important research context because they equalise access by design. When a public agency delivers training, platform access and institutional support to a cohort of firms, the resulting heterogeneity in outcomes cannot be attributed to differential access. It remains an empirical question. This makes such programmes useful settings for studying value capture and outcome inequality.

In Malaysia, the Federal Agricultural Marketing Authority (FAMA) introduced the Agromarketing Masterclass TikTok Shop Edition (AMTTSE) as a structured intervention to equip agropreneurs with practical digital marketing, livestreaming and content creation capabilities. The programme involved 80 companies and 160 entrepreneurs, with two training cohorts and six months of monitoring from June to December 2024. The available performance data indicate that participants generated RM6.2 million in cumulative sales over the programme period (FAMA, 2024).

The empirical puzzle is that substantial aggregate GMV coexists with substantial seller-level concentration and inactivity. The top 10 sellers accounted for 52.4 per cent of December GMV, the seller-level Gini coefficient was approximately 0.71, and 25 of 80 companies recorded zero December GMV. Aggregate success and seller-level inequality are not contradictory; they are two faces of the same phenomenon, and understanding their relationship is the central task of this article.

Three research questions guide the analysis. RQ1 asks how concentrated commercial value capture is among firms participating in a common government-supported social-commerce programme. RQ2 asks how content-led and transaction-oriented TikTok Shop mechanisms correspond with programme-level commercial performance. RQ3 asks what heterogeneous value capture reveals about the relationship between institutional access, platform participation and firm-level capability.

The article makes four contributions. First, it provides a verified seller-level concentration analysis of a government-platform social-commerce training programme, using administrative GMV records that permit a direct test of the access-conversion gap. Second, it adds a distributional lens to research on digital entrepreneurship interventions. Third, it supplies a low-cost inclusion dashboard that any agency collecting a GMV sheet can implement. Fourth, it contributes to the entrepreneurship and strategic management literature by shifting the analytical lens from adoption to distribution.

The remainder of the article proceeds as follows. Section 2 reviews the literature and develops the theoretical framework. Section 3 describes the research context and methodology. Section 4 presents the results. Section 5 discusses the findings. Section 6 articulates the theoretical contribution, Section 7 the managerial and policy implications, Section 8 the limitations and Section 9 the conclusion, followed by references.

2.1 Social commerce and platform-enabled entrepreneurship

Social commerce is widely defined as the use of social media platforms to support commercial transactions, distinguished from conventional e-commerce by the centrality of social interaction, user-generated content and community engagement (Hajli, 2015; Busalim & Hussin, 2016). The more recent livestreaming literature demonstrates that real-time video demonstration builds trust and engagement in ways that static listings cannot (Wongkitrungrueng & Assarut, 2020; Sun et al., 2019). For platform-enabled entrepreneurship, the implication is that content-driven formats are commercially relevant, but the present study does not measure the specific behavioural mechanisms behind seller performance.

2.2 Value creation versus value capture

Value capture theory sharpens the distinction between value creation and value appropriation: firms differ in their capacity to appropriate the value their activities create (Lepak, Smith & Taylor, 2007; Gans & Ryall, 2017). The platform economy introduces a distinctive distributional dynamic: platform owners control visibility, ranking and algorithmic distribution, while sellers compete for attention within rules they do not set (Gawer, 2014; Jacobides, Cennamo & Gawer, 2018). This asymmetry creates platform-dependent entrepreneurs (Cutolo & Kenney, 2021). Performance inequality within a single programme is the empirical footprint of this dynamic.

2.3 Digital access and observed outcomes

Public agencies across emerging economies use training programmes, subsidies and platform partnerships to improve market access for small producers (Barrett, 2008; Reardon & Timmer, 2014). The digital-inequality literature formalises this as a multi-level gap: access, skills and usage, and realised outcomes (van Deursen & van Dijk, 2019; Scheerder et al., 2017). The present dataset does not directly observe the mechanisms linking access to realised outcomes, so the defensible focus is on distributional outcomes rather than mechanism testing.

2.4 Research questions

Three research questions guide the analysis. RQ1 asks how concentrated commercial value capture is among firms participating in a common government-supported social-commerce programme. RQ2 asks whether aggregate programme totals mask uneven value capture. RQ3 asks how sensitive programme-level performance is to the best-performing sellers. These questions are narrower than a general programme evaluation and can be answered using the administrative record without inventing causal mechanisms.

3.1 Research design

The study adopts a retrospective secondary analysis of programme-administrative data. The design is descriptive and distributional in intent: it measures how value is spread across sellers and assesses how sensitive the aggregate picture is to leading sellers. No claim of causal inference is made from this design.

3.2 Research context

AMTTSE provides a bounded empirical setting in which agropreneurs engaged with a common training and platform environment. Administrative performance records span June-December 2024. The programme involved 80 companies and 160 entrepreneurs, with two training cohorts and six months of monitoring.

3.3 Sample and population

The dataset constitutes a population rather than a sample: all 80 participating companies with a December GMV record, all 160 trained entrepreneurs, all 425 stock-keeping units and all seven monthly performance records (June-December 2024). Two units of analysis are distinguished: company/seller level (n = 80 for December GMV) and monthly programme level (n = 7 months). Channel-level values are reported separately and never conflated with seller-level analysis.

3.4 Data source

The dataset used is the Projek Perintis Report (December 2024), consisting of two sheets: Summary and Seller GMV. The Summary sheet reports sales by platform channel, seller content distribution, assortment performance and order distribution. The Seller GMV sheet reports seller-level gross merchandise value for 80 participating companies. The programme period used for main analysis is June to December 2024.

3.5 Measures

Variables were defined operationally before analysis: GMV, active seller, zero-GMV seller and channel share. Concentration measures include the Gini coefficient, Lorenz curve, top-k shares and the bottom-50 per cent share.

3.6 Analytical procedures

The Gini is computed on the full 80-seller distribution including the 25 zero-GMV sellers, because excluding them would understate inequality by treating non-participation as non-existence. Robustness is assessed through the active-seller subset and sensitivity checks that remove leading sellers from the total. Statistical methods were selected for their appropriateness to skewed, non-normal seller-level distributions.

3.7 Research ethics and data governance

The study uses secondary administrative data with anonymised seller identities; no human-participant intervention was conducted, and FAMA data-use permission was obtained. Because the data contain no personal identifiers and were collected for programme monitoring rather than research, the ethical risk is low.

3.8 Data availability

The data supporting the findings of this study are available from the corresponding author upon reasonable request and with the permission of FAMA.

5. Discussion

Aggregate programme success and participant-level value capture are not the same thing. AMTTSE achieved substantial total sales, but realised gains were highly concentrated among a small number of sellers. The data do not identify why this concentration occurred. Several mechanisms remain plausible, including prior brand strength, product attractiveness, advertising intensity, seller effort, inventory availability, fulfilment capacity, and algorithmic exposure. None of these mechanisms can be isolated from the current administrative record, so they remain future research questions rather than findings.

The policy implication is straightforward: agencies should report aggregate GMV together with distribution metrics, including active-seller rate, zero-GMV rate, median GMV, top-k shares, and the Gini coefficient. Such reporting makes it harder for headline totals to mask uneven realised outcomes.

Key limitations are straightforward. First, the study is a secondary analysis of one programme and one platform, so transferability is limited. Second, the design is descriptive rather than causal. Third, the administrative record does not contain direct measures of capability, brand strength, advertising expenditure, inventory, fulfilment capacity or algorithmic exposure. Fourth, GMV is revenue-related transaction value, not profit. Fifth, the programme-level series contains only seven monthly observations.

Download the complete GamaIJB submission gamaijb-AMTTSE-New-Submission.docx — a submission-ready manuscript (5,500-6,500 words) following the GamaIJB author guidelines in full: title, abstract, keywords, JEL classification, introduction, literature review, methods, results, discussion, conclusion, limitations, references in APA style, six tables and three figures.

Funding statement

This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors. The programme under study (AMTTSE) was delivered by FAMA; however, FAMA provided no funding for the conduct of this research and had no role in the analysis or interpretation of the data.

Ethics statement

This study analysed anonymised secondary administrative data collected by FAMA for programme monitoring purposes. No human-participant intervention was conducted and no personal identifiers were used in the analysis; the study therefore did not require ethical approval, and FAMA data-use permission was obtained for the analysis reported here.

Conflict of interest statement

The authors declare that they have no conflict of interest. [AUTHOR TO CONFIRM]

Data availability statement

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request and with the permission of FAMA.

Author contributions statement

[AUTHOR TO CONFIRM]

Acknowledgements

[AUTHOR TO CONFIRM]

Prior publication disclosure

Two earlier accounts of the AMTTSE programme were previously published: an aggregate programme-level evaluation in the Journal of International Business, Economics and Entrepreneurship (Abd Razak et al., 2026a; DOI 10.24191/jibe.v11i1.11085) and a capability-based study in the Journal of Advanced Research in Business and Management Studies (Abd Razak et al., 2026b; DOI 10.37934/arbms.44.1.1032). The present manuscript constitutes a new distributional treatment of the same dataset: it addresses realised value capture rather than programme effectiveness or capability relationships, and it uses seller-level concentration analysis, top-k shares and sensitivity checks. The authors confirm that this manuscript has not been published elsewhere and is not under consideration by any other journal.

AI-assisted technology disclosure

The authors used AI-assisted writing technology to assist with language refinement, referencing and formatting of this manuscript in line with GamaIJB guidance on the use of AI. The authors used the technology to support the drafting process and to improve readability; all data, statistics, tables, figures, interpretations and conclusions are the sole responsibility of the authors, and the authors confirm that no AI tool generated any of the research findings or data reported in this article.

ORCID and CRediT

ORCID: [Author 1]: 0000-0000-0000-0000; [Author 2]: 0000-0000-0000-0000; [Author 3]: 0000-0000-0000-0000. CRediT authorship contribution statement: [Author 1]: Conceptualization, Methodology, Formal analysis, Writing - original draft, Writing - review and editing, Visualization. [Author 2]: Supervision, Validation, Writing - review and editing. [Author 3]: Investigation, Data curation, Writing - review and editing.