After the demonstrated success of UPI in payments, the Open Network for Digital Commerce (ONDC) was conceived as the next population-scale digital public infrastructure: an open, interoperable protocol that would unbundle e-commerce the way UPI had unbundled payments. By 2026, ONDC's transaction volume is a small share of the e-commerce market, while the closed platforms it was designed to challenge have grown further. The protocol works as designed; the adoption does not follow. What did the architecture not anticipate?
The Open Network for Digital Commerce was conceived as a transformative intervention. The logic was elegant. Just as the Unified Payments Interface unbundled payments from specific banking apps, ONDC would unbundle e-commerce by separating the buyer app from the seller app. The network would be open, the protocol standardised, and no single platform would control discovery, pricing, or logistics.
The problem is that e-commerce is not payments. UPI succeeded because a payment is a simple, standardised transaction: move a defined amount from account A to account B. ONDC attempted to protocol-ise a market where product discovery, catalogue quality, delivery logistics, returns management, customer trust, and post-sale service are each complex, capital-intensive operational challenges that closed platforms spent billions of dollars building over a decade.
The catalogue problem surfaced first. Sellers listed products with inconsistent descriptions, poor images, and no reviews. The protocol ensured interoperability. It could not ensure quality.
Logistics revealed the second fault line. A buyer ordering from a small seller in another city faced delivery timelines and returns processes visibly inferior to the platform experience. The protocol connected buyer to seller. It did not solve the physical infrastructure between them.
Trust and grievance redressal created the third barrier. When something goes wrong on ONDC, the accountability chain fragments across buyer app, seller app, logistics provider, and seller. The network architecture designed to prevent platform dominance also prevented the concentration of accountability that consumers rely on.
ONDC was set up as a Section 8 company with Department for Promotion of Industry and Internal Trade (DPIIT) oversight, but it lacked regulatory authority to mandate participation by large platforms. Without mandated interoperability, it operated as a voluntary network competing with incumbents who had no commercial incentive to join a protocol that would commoditise their competitive advantages. Protocol-layer interventions work when the underlying transaction is standardised and the infrastructure requirements are minimal; UPI met both conditions and e-commerce meets neither.
The contrast with UPI's success is instructive beyond the technical dimension. Two payment platforms together control roughly 85% of UPI transactions; the very platform concentration in payments that ONDC was designed to prevent in commerce. UPI succeeded because the transaction is standardised, the infrastructure requirement is minimal, and the protocol could ride on existing banking rails. ONDC attempted to create an equivalent protocol for commerce, where the transaction involves discovery, logistics, returns, trust, and grievance redressal, each requiring operational infrastructure that no protocol can substitute.
Connecting buyer to seller is the easy part; delivering the product, resolving the dispute, and earning the trust is the hard part; no protocol solves that.