The Working Capital Paradox
India's MSME sector faces ₹40 lakh crore working capital gap yearly. Banks don't lend below 48-month tenure. NBFCs charge 18-24% for invoice discounting. Yet adoption remains painfully low—under 5% of eligible invoices get discounted.
Why? Because one-off transactions feel expensive and frictionless.
A garment exporter with ₹50 lakh monthly receivables pays ₹75,000 per transaction. That's 1.5% per draw. If he borrows 12 times yearly, he's paying 18% annualized. But he only does it when desperate.
From Transaction to Subscription
The best platforms in Southeast Asia (Singapore, Indonesia) shifted to subscription models. OKCash in Indonesia charges ₹800-1200/month. Fundbox in US charges $99-149/month for unlimited draws at 1-3% rates.
Margins math: ₹1000/month × 1000 customers = ₹10 lakh recurring. Per-transaction model: 200 transactions/month at 0.5% each on ₹50 lakh AUM = ₹5000 revenue per transaction, ₹10 lakh monthly—same outcome, but unpredictable.
Subscription wins because: (1) Predictable cash flow for lending. (2) Customer doesn't optimize to use you less. (3) Retention becomes survival metric.
The Churn Cliff
Horizontal SMB platforms in India see 12-18% monthly churn. A customer borrows in month 1, avoids month 2 due to better working capital, churns in month 3.
Embedded platforms perform differently. Kredent (embedded in GST invoice flow) reports <3% annual churn. Why? The invoice discounting isn't a separate step—it's native to their workflow.
This resembles SaaS retention dynamics. A standalone expense tracking tool has 40% annual churn. Embedded inside accounting software, it's 5-10%. The user never leaves because they're not using "invoice discounting"—they're using a ledger.
Net Revenue Retention Above 110%
NRR above 110% in this space requires vertical concentration. Textiles, auto-parts, pharmaceuticals—industries with predictable payment cycles and repeat invoicing.
A horizontal fintech targeting ₹50-200 lakh revenue businesses will see NRR of 95-105%. Why? As the business scales, it graduates to bank credit. You become the training wheels. Losing tier-1 customers is inevitable.
A vertical platform (say, textile merchants only) locks supply chain dynamics. Supplier A always pays Supplier B on day-45. Merchant C needs money on day-10. The platform becomes the payment choreographer. NRR reaches 115-125% from same-customer expansion and deeper penetration.
The Subscription vs Usage Hybrid
Optimal structure: ₹1000/month flat + 0.25-0.5% discount rate, not 1-1.5%.
This shifts psychology. The customer pre-pays for the relationship (₹12,000/year). She then uses discounting only when it beats her next-best alternative. Platform gets predictable base revenue + volume upside.
Customer LTV calculation flips: instead of (discount rate × AUM × tenure), it becomes (monthly subscription + (discount rate × incremental usage volume) × tenure).
For a ₹10 crore annual turnover textile exporter: ₹1000/month + 0.5% on ₹50 lakh average draw = ₹1000 + ₹2500 = ₹3500/month. Annualized revenue: ₹42,000. CAC payback in 3-4 months if acquisition costs ₹8-10K.
India Stack + Timing
GST integration and Udyam registration have improved SMB formalization. Invoice digitization is now possible at scale. This wasn't true in 2015.
But competition is intensifying. HDFC Bank, ICICI Bank, and 50+ fintech platforms now offer invoice discounting. Unit economics are compressing.
The winner won't be the cheapest lender. It'll be whoever locks the recurring relationship first. Before churn becomes the ceiling.
The Build Decision
If you're starting in this space: (1) Pick a vertical with <10 SKU invoice profiles. (2) Build your subscription plan first, pricing second. (3) Make API integration free and mandatory. (4) Track NRR obsessively from month 6 onward.
Transaction-based fintech platforms in India are racing toward ₹0 unit economics. Subscription-embedded fintech will own SMB relationships for the next decade.