Ecosystem Description in Platform Company Prospectuses: Network Effect Assessment

The SFC’s 2024-25 enforcement priorities, as outlined in its annual report published in June 2025, specifically flagged “misleading descriptions of business ecosystems” in IPO prospectuses as a key area of concern, with two unnamed platform companies receiving warning letters regarding their network effect claims. This regulatory scrutiny arrives as the HKEX processes a backlog of at least 12 platform-company listing applications in Q3 2025, each relying heavily on ecosystem narratives to justify valuation premiums. The problem is structural: network effects are notoriously difficult to quantify, yet they underpin the pricing of companies from food delivery to fintech. When a prospectus states a platform connects “X million merchants to Y million consumers,” the market needs a framework to assess whether that connection creates genuine, defensible value or merely a crowded digital marketplace. This article establishes a methodology for evaluating ecosystem descriptions in HKEX Main Board and GEM prospectuses, drawing on the SFC’s Code of Conduct for Corporate Finance Advisors (Chapter 17) and the HKEX Listing Rules’ requirements for “forward-looking statements” under Chapter 11.

The Anatomy of Network Effect Claims in Prospectuses

Direct vs. Indirect Network Effects: The SFC’s Implicit Distinction

The SFC has not issued a standalone circular on network effects, but its 2024 thematic review of technology company IPO disclosures (published 15 October 2024) provided a de facto classification. The review examined 18 prospectuses filed between January 2022 and December 2023, finding that 14 claimed “network effects” without distinguishing between direct and indirect types. This conflation is material because the two categories carry different competitive durability profiles. Direct network effects — where each additional user increases the value of the service for all other users, as seen in messaging platforms like WeChat — create strong lock-in. Indirect network effects — where more users on one side attract more participants on another side, typical of e-commerce marketplaces — are more fragile, as participants can multi-home across platforms.

A HKD 8.2 billion Main Board listing application from a Southeast Asian food delivery platform filed in March 2025 illustrates the problem. The prospectus’s “Competitive Strengths” section (pages 45-52) stated: “Our platform benefits from powerful network effects, as more diners attract more restaurants, and more restaurants attract more diners.” This is a textbook indirect network effect. However, the prospectus did not quantify the multi-homing rate — the percentage of restaurants also listed on competing platforms GrabFood or Deliveroo. The SFC’s 2024 review noted that only 3 of the 18 prospectuses provided multi-homing data. Without this metric, an investor cannot assess whether the claimed network effect is proprietary or merely a function of overall market growth.

The “Active User” Definition Trap

HKEX Listing Rule 11.07 requires that any operating metric used in a prospectus be “clearly defined and consistently calculated.” Yet ecosystem descriptions routinely deploy “active users” as a proxy for network strength without specifying the denominator. A 2025 GEM listing application for a Hong Kong-based pet-sitting platform defined “monthly active users” as “any user who logged into the app at least once in the preceding 30 days.” This definition inflates engagement counts by including users who may have logged in only to check a single notification — a practice the SFC’s 2024 review labelled “potentially misleading” in its observations on pages 34-36.

The more rigorous approach is to require a “core user” definition that captures transaction-initiating activity. For the pet-sitting platform, a core user would be a pet owner who completed at least one booking or a sitter who accepted at least one request in the period. The difference is material: the platform disclosed 245,000 monthly active users in its prospectus summary, but a back-of-envelope calculation using the disclosed transaction data (142,000 completed bookings per month) suggests a core user count of approximately 85,000 to 100,000, assuming each core user completes 1.4 to 1.7 transactions per month. This 58-65% discrepancy would directly affect any network effect multiplier used in valuation.

Quantifying Ecosystem Depth: Beyond User Counts

Transaction Density and Cross-Side Conversion Rates

The HKEX’s 2023 Guidance Letter GL113-23 on “Disclosure of Non-Financial KPIs” (updated January 2024) explicitly encourages issuers to present “density metrics that demonstrate the depth of interaction” within a platform ecosystem. The most useful metric is cross-side conversion rate: the percentage of users on one side who convert into transacting with the other side within a defined period. For a ride-hailing platform, this would be the percentage of passengers who complete at least one trip per month divided by the total registered passengers.

A 2024 Main Board prospectus from a Chinese electric vehicle charging platform provided a rare example of best practice. On page 78, it disclosed that 68.2% of registered charging station operators had received at least one charging session request from a registered EV owner in the preceding quarter, and that 41.7% of those requests resulted in a completed session. This two-layer conversion rate — cross-side connection followed by transaction completion — allows an analyst to calculate a “platform efficiency ratio” of 28.4% (68.2% × 41.7%). Compare this to a competitor’s prospectus that simply stated “over 1 million charging sessions completed in 2024” without any conversion context. The first prospectus enables a network effect assessment; the second does not.

The “Density Threshold” Problem in GEM Listings

GEM-listed platform companies face a particular challenge: they often lack the scale to demonstrate network effects convincingly. The GEM Listing Rules (Chapter 16) require a minimum market capitalisation of HKD 100 million at listing, but no specific requirement for platform density. A 2025 GEM prospectus for a Hong Kong-based freelance services marketplace disclosed 12,000 registered freelancers and 8,000 registered clients, with 4,200 completed projects in the preceding 12 months. The implied “projects per freelancer” ratio of 0.35 per year suggests most freelancers have not found work through the platform — a weak network effect at best.

The SFC’s 2024 review implicitly addressed this by recommending that GEM issuers disclose “the percentage of registered participants on each side who have completed at least one transaction in the preceding 12 months.” For the freelance marketplace, this would have been approximately 35% for freelancers (4,200 projects divided by 12,000 freelancers) and 52.5% for clients (4,200 projects divided by 8,000 clients). These figures, once disclosed, would allow an investor to assess whether the platform is a genuine two-sided marketplace or a glorified directory service. The prospectus instead buried the project count in a footnote on page 112, while highlighting the 12,000 and 8,000 registration numbers in the summary section — a classic “vanity metric” approach.

Multi-Homing and Switching Costs: The Competitive Moat Test

Measuring Multi-Homing Rates from Public Filings

The most direct test of a platform’s network effect durability is its multi-homing rate — the proportion of participants on either side who also use competing platforms. HKEX Listing Rule 11.08 requires disclosure of “material competitive factors,” but does not explicitly mandate multi-homing data. A 2024 prospectus from a Hong Kong-based food delivery platform was unusual in voluntarily disclosing that 62% of its restaurant partners also listed on at least one competitor, and that 48% of its active consumers had used a competing app in the preceding month. These figures, disclosed on page 91, effectively neutralised the prospectus’s earlier claim of “exclusive ecosystem lock-in.”

An analyst can approximate multi-homing rates from disclosed data even when not explicitly provided. For a ride-hailing platform that discloses total trips, active drivers, and active riders, the ratio of trips per driver per day can indicate multi-homing. If a driver completes 2.1 trips per day on Platform A, but the market average is 4.5 trips per day across all platforms, the implication is that the driver is multi-homing on at least one other platform. The SFC’s 2024 review flagged this analytical approach as “acceptable supplementary analysis” in its observations on page 42, though it cautioned that such estimates must be clearly labelled as assumptions.

Switching Costs as a Structural Barrier

Network effects are only valuable if they create switching costs. The HKEX’s 2023 Guidance Letter on “Business Model Disclosure” (GL112-23) recommends that issuers “explain the mechanisms by which participants become locked into the platform.” Common mechanisms include data accumulation (e.g., a restaurant’s menu and order history), reputation systems (e.g., seller ratings that cannot be transferred), and payment integration (e.g., stored wallets).

A 2025 Main Board prospectus for a cross-border payment platform provided an exemplary switching cost analysis on pages 56-60. It disclosed that the average merchant had processed 347 transactions on the platform, with an average stored balance of HKD 12,800. It then calculated a “switching friction index” of 2.4 months of transaction value — meaning a merchant would need to forgo 2.4 months of processing revenue to migrate to a competitor. This index was derived from the time required to re-integrate payment APIs (average 45 days), re-verify compliance documentation (average 18 days), and rebuild transaction history for credit underwriting (average 10 days). The prospectus then benchmarked this index against the industry average of 1.1 months, providing a quantitative basis for its claim of “superior ecosystem stickiness.”

Regulatory and Enforcement Implications

The SFC’s Emerging “Ecosystem Disclosure” Standard

The SFC’s 2024-25 enforcement priorities explicitly mention “misleading ecosystem descriptions” as a subset of false or misleading statements under Section 277 of the Securities and Futures Ordinance (Cap. 571). In a 2024 enforcement action against a now-delisted e-commerce platform, the SFC alleged that the prospectus had claimed “over 1 million active merchants” when internal data showed that only 340,000 merchants had completed a transaction in the preceding six months. The case, settled in December 2024 with a HKD 15 million fine and a two-year director disqualification, established a precedent: “active” in an ecosystem context must mean “transaction-active,” not “registered.”

Sponsors and reporting accountants should expect the SFC to request the following in any platform company IPO filing: (1) a breakdown of registered vs. transaction-active participants on each side of the platform, (2) multi-homing rates for at least one side, supported by survey or third-party data, and (3) a switching cost analysis with quantified time and monetary components. The HKEX’s Listing Committee, in its 2025 Annual Review (published March 2025), confirmed that it would “require enhanced ecosystem disclosures” for all platform company applications effective 1 July 2025, with specific reference to the metrics outlined above.

The “Network Effect Multiplier” in Valuation: A Disclosure Gap

A persistent issue is the use of network effect multipliers in valuation sections of prospectuses. A 2024 Main Board prospectus for a social commerce platform included a valuation analysis (pages 120-125) that applied a “network effect premium” of 1.8x to its discounted cash flow valuation, without defining how the multiplier was derived. The SFC’s 2024 review noted that such multipliers “lack empirical support and may mislead investors” (paragraph 4.7). The review recommended that any multiplier used in a valuation be derived from a peer group analysis with clearly stated comparables, and that the sensitivity of the valuation to changes in the multiplier be disclosed.

For the social commerce platform, a sensitivity analysis would have shown that reducing the multiplier from 1.8x to 1.2x — still above the market average for non-platform companies — would reduce the implied valuation by HKD 4.2 billion, or 33% of the proposed market capitalisation. The prospectus did not disclose this sensitivity. The HKEX’s updated Guidance Letter on “Valuation Disclosures in Listing Documents” (GL56-23, revised April 2025) now explicitly requires such sensitivity analysis for any non-financial valuation input, including network effect multipliers.

Actionable Takeaways

1. Demand a “transaction-active” definition for every ecosystem participant category in a prospectus, and calculate the ratio of transaction-active to registered participants for each side of the platform — a ratio below 40% for the supply side typically indicates weak network effects.
2. Require disclosure of multi-homing rates for at least one side of the platform, using either survey data or third-party analytics, and cross-reference this against the prospectus’s “competitive moat” claims — a multi-homing rate above 50% on either side effectively negates exclusivity-based network effect arguments.
3. Insist on a switching cost analysis that quantifies both time (in days) and monetary cost (as a percentage of average participant revenue), and benchmark these figures against industry averages disclosed in comparable prospectuses or SFC guidance letters.
4. Verify that any network effect multiplier used in valuation is derived from a documented peer group analysis, with sensitivity tables showing the impact of a 0.5x change in the multiplier on the implied market capitalisation.
5. Check that the prospectus’s “ecosystem” description is consistent across all sections — a common red flag is when the “Business Overview” section (typically written by the issuer) claims proprietary network effects, while the “Risk Factors” section (typically drafted by the sponsor) warns that participants can easily switch to competitors.