INTERFACTORY INC

1. Company Overview

INTERFACTORY INC. is a Japanese cloud-commerce software company listed on the Tokyo Stock Exchange Growth market under code 4057. Its core product, EBISUMART, supports the construction and operation of customized e-commerce sites. The company is small: the supplied market snapshot values its equity at ¥1.445 billion, or ¥358 per share, as of 1 September 2026.

The latest clean official annual base is FY2026, ended 31 May 2026. Its annual securities report was filed on 19 August 2026, after the 13 July preliminary annual earnings release. The FY2026 figures below are therefore audited annual data: revenue was ¥2.862 billion, operating profit ¥64.7 million, net income ¥34.3 million, operating cash flow ¥145.5 million, capital expenditure ¥208.3 million, and net assets ¥1.212 billion. FY2027 guidance, announced on 13 July 2026, is a company guidance / management update, not an official result: revenue ¥2.977 billion, operating profit ¥111 million, and net income ¥72 million.

Net cash or net debt is unavailable in the supplied deterministic snapshot and must not be treated as zero. The FY2026 annual report does show ¥386.7 million of cash and ¥350.0 million of short-term borrowings, but that is not a complete reconciled net-debt measure; the snapshot specifically lacks the short-term-securities input needed for its net-cash calculation. The balance sheet is therefore adequate for assessing liquidity, but not for making a precise net-cash valuation adjustment.

The central issue is not survival. Equity represented 60.8% of FY2026 assets, and the company remained profitable. The issue is whether EBISUMART’s recurring revenue base can earn attractive returns after funding product development, sales capacity, AWS infrastructure, and newer businesses. FY2026 was a sharp profit reset after a strong FY2025 recovery.

2. How the Company Makes Money

INTERFACTORY has three reported businesses. The economically important one is the cloud-commerce platform business, built around EBISUMART, EBISUMART Lite, BtoB, and Enterprise. It produces an initial flow of customization and implementation revenue, followed by recurring monthly operation and maintenance fees. Those recurring fees include fixed charges, usage-linked charges, and optional-service fees. The company also earns from further customization after go-live.

In FY2026, the cloud-commerce platform business generated ¥2.732 billion of revenue, or roughly 95% of company revenue. Within it, system operation and maintenance revenue was ¥1.747 billion, up 6.4% year on year; system-development revenue was ¥961.7 million. This recurring maintenance stream is the franchise’s most valuable component. Management reported monthly average ARPU of ¥416 thousand, up from ¥384 thousand in FY2025, and GMV processed through the platform of ¥204.5 billion, up from ¥169.7 billion. These are official management KPIs disclosed in the FY2026 annual report, not independently audited industry data.

The other two businesses are much smaller and currently dilute group economics. EBISU GROWTH provides consulting, operations, marketing, and fulfillment support to e-commerce merchants. Its FY2026 revenue fell 47.5% to ¥129.7 million because advertising operations were shifted to a partner and were thereafter reported on a net-fee basis rather than gross revenue; its segment result nevertheless turned slightly positive at ¥4.4 million. The data-utilization platform business, centered on EBISU PIM, recorded only ¥0.7 million of revenue while producing a ¥59.0 million segment loss.

This is not a low-capital-intensity SaaS model in the pure sense. Physical capital needs are modest, but software development is capitalized. FY2026 capital expenditure was ¥208.3 million, principally ¥143.4 million for EBISUMART and Enterprise functionality and ¥58.6 million for EBISU PIM. Year-end software assets were ¥616.1 million, about half of net assets. This accounting treatment defers some economic cost into future depreciation. FY2026 operating cash flow less total capital expenditure was negative ¥62.8 million; this is simple arithmetic from audited cash-flow data, not a company-reported free-cash-flow metric.

The franchise has real switching costs: an operating e-commerce system is integrated with payment, inventory, logistics, ERP, and merchant workflows, and migration is disruptive. Its partner ecosystem and more than 800 stated EBISUMART customers add some practical friction. But it is not yet demonstrated to be a true high-return compounder. The company itself states that technological barriers to entry are not especially high. Revenue has grown, but profitability has been volatile and incremental investment has not yet produced consistently superior owner returns.

3. Why the Stock Fell

The supplied research does not include a verified share-price history, prior peak, or dated drawdown. It is therefore not possible to state the exact magnitude or timing of a decline, or to attribute every price movement to one announcement. The evidence does support a clear deterioration in market-relevant operating expectations.

The key event was the 13 July 2026 FY2026 result. INTERFACTORY had guided in July 2025 for FY2026 revenue of ¥3.223 billion, operating profit of ¥82 million, and net income of ¥53 million. Actual FY2026 revenue was ¥2.862 billion, operating profit ¥64.7 million, and net income ¥34.3 million. That was a shortfall of ¥361 million in revenue, ¥17 million in operating profit, and ¥19 million in net income versus the company’s own prior guidance.

The miss had three identifiable components. First, Enterprise development was delayed after a review of the development structure, causing the core cloud-commerce business to deliver ¥2.732 billion of revenue versus the company’s ¥3.001 billion plan. Second, EBISU PIM took longer to gain market recognition: revenue was only ¥0.7 million against a ¥49 million plan while its segment loss widened to ¥59.0 million. Third, new system-development order intake weakened, reducing the pipeline that normally feeds future recurring maintenance revenue.

The market narrative is therefore understandable: an expensive product-investment cycle has arrived before evidence that Enterprise and PIM can be commercialized, while the legacy business has lost some new-order momentum. The problem is not that recurring revenue collapsed in FY2026; maintenance revenue grew. The problem is that the earnings engine below that revenue base is too weak to validate a growth-company valuation unless new-product investments and new order intake recover.

4. What the Market Is Assuming

At a ¥1.445 billion market capitalization, the market appears to be assigning little value to a rapid return to FY2025 profitability. It is pricing the company more like a modest recurring-software franchise with impaired returns on incremental investment than like a high-growth platform. That interpretation is reasonable, but it is not conclusively correct.

The multi-year record explains the discount. Audited revenue rose from ¥2.170 billion in FY2021 to ¥2.862 billion in FY2026, but operating profit moved from ¥208.6 million to ¥36.4 million, ¥53.3 million, negative ¥24.9 million, ¥196.3 million, and then ¥64.7 million. The supplied deterministic research calculates a 5.69% revenue CAGR but a negative 23.64% net-income CAGR. In other words, sales growth has not translated into stable compounding of owner earnings.

FY2027 guidance assumes improvement, not a full restoration: revenue is projected to rise 4.0% and operating profit 73.0% to ¥111 million. Management also expects maintenance revenue to decline 6.7%, reflecting churn, while expecting price revisions to support the revenue base and system-development revenue to recover. This is a meaningful tension: the company’s main recurring engine is forecast to face churn pressure, while the recovery rests partly on improved new development sales. There is no independent evidence yet that this recovery is secured.

5. Temporary or Structural?

Overall diagnosis: structural but survivable. FY2026’s earnings decline is not evidence that the existing EBISUMART customer base has failed. The maintenance, GMV, and ARPU data point the other way. But the company’s weak barriers to entry, reliance on skilled labor, capitalized development, and uncertain new-product economics are structural limitations on returns. The repairable part is revenue execution; the non-repairable risk is that the business never earns sufficient returns on the capital required to stay relevant.

The most important positive counterweight is the existing recurring base. A customer that has customized its commerce workflow is not likely to move merely because another vendor offers a similar feature. The most important negative counterweight is that this switching friction does not automatically confer pricing power or high returns. INTERFACTORY must continually spend to maintain functionality, support integrations, meet security standards, and prevent better-capitalized rivals from eroding its relevance.

6. Is the Market Wrong? By How Much?

The market is not obviously wrong on the current evidence. It is correctly discounting weak and volatile reported profits, uncertain new-product monetization, and low stated technological entry barriers. It may be underestimating the durability of the recurring maintenance base and the possibility that capitalized investment ultimately supports higher owner earnings. But that is a conditional upside case, not a demonstrated mispricing.

Moat-and-mispricing score: 4/10 overall. The moat is 3/10: customer switching costs and installed-base economics are real, but technological barriers are modest and returns are not stable. The mispricing score is 5/10: the market price embeds skepticism, yet the supplied owner-earnings model creates substantial upside if its definition proves economically valid. The uncertainty around that definition prevents a higher score.

Worth-dead evidence is limited. The supplied deterministic snapshot estimates NCAV at ¥451.0 million, only 31.22% of the current market capitalization. Liquidation value is unavailable, and net cash is unavailable. This is therefore not a classic net-net or balance-sheet-protected cigar butt. The supplied book value is ¥1.211 billion and price-to-book is 1.19x, but much of book value is not hard liquidation value because FY2026 software assets were ¥616.1 million.

Worth-alive evidence is more attractive but less reliable. The supplied deterministic model estimates owner earnings at ¥305.4 million, implying a 21.14% owner-earnings yield and a 4.73-year payback at the 1 September 2026 market capitalization. These are model estimates, not audited facts. They conflict materially with FY2026 audited net income of ¥34.3 million, operating cash flow of ¥145.5 million, and negative cash flow after total capital expenditure. The supplied 413.5% free-cash-flow conversion figure signals that the model’s treatment of sustaining development expenditure differs sharply from reported cash investment. Without the model’s full maintenance-capex and capitalization assumptions, ¥305.4 million should be treated as an upside cross-check, not as base owner earnings.

The following is a FY2026-based intrinsic-value framework, using FY2027 guidance only as a reference point. It capitalizes post-interest, post-tax normalized owner earnings directly into equity value. No separate net-cash/debt adjustment is applied because net cash/debt is unavailable in the authoritative snapshot. This is intentional: adding an unverified cash adjustment would create false precision. Values are intrinsic-value estimates, not price targets.

The valuation bridge is simply normalized owner earnings divided by the required equity yield: ¥80 million / 14.0% = ¥571 million; ¥160 million / 12.0% = ¥1.333 billion; and ¥250 million / 10.5% = ¥2.381 billion. Per-share values use 4.036 million shares outstanding excluding treasury shares, from the FY2026 earnings disclosure. The central range is therefore roughly ¥0.57 billion to ¥2.38 billion of equity value, or ¥142 to ¥590 per share.

As a sensitivity only, mechanically capitalizing the supplied ¥305.4 million owner-earnings model at a 10.5% required yield would imply about ¥2.91 billion of equity value, or about ¥721 per share. That would be roughly 101% above the supplied market price. It is not the base case because the model’s owner-earnings definition has not been reconciled to reported cash investment. The market is probably pricing a normalized earnings level between the bear and base cases; it is not pricing a sustainable ¥250 million to ¥305 million owner-earnings outcome.

7. Key Facts, Estimates, and Judgments

The strongest falsification check is the recurring base. The thesis becomes materially worse if maintenance revenue, ARPU, GMV, or retention weaken simultaneously, because that would show that the franchise rather than merely the new-product investment is impaired. A second check is cash economics: the investment case fails if software additions remain near FY2026 levels without a sustained improvement in operating cash flow, because reported earnings would then be overstated relative to owner earnings.

Evidence that would improve the thesis would include two consecutive periods of recovering development order intake, stable or rising maintenance revenue despite disclosed churn, meaningful PIM revenue with narrowing losses, and cash flow after software investment turning consistently positive. Evidence that would change the thesis negatively would include a software impairment, persistent churn that defeats pricing actions, a security or availability failure, a further decline in development orders, or FY2027 results materially below the ¥111 million operating-profit guidance.