ONDC vs. Amazon: What Actually Happened When India Built Its Own E-commerce Protocol

India launched ONDC — the Open Network for Digital Commerce — with an explicit ambition: break Amazon and Flipkart's duopoly by building an interoperable e-commerce infrastructure that any seller or buyer app could plug into.

It's been running at scale for about two years now. The results are more complicated than either its promoters or critics want to acknowledge.

## What ONDC Actually Is

ONDC is not an app. It's a protocol — think of it as the equivalent of UPI but for commerce rather than payments. The idea is that instead of every seller needing to list on Amazon or Flipkart to reach buyers, they can list on any ONDC-compatible "seller app" and be discoverable to buyers using any "buyer app."

The structural argument for why this matters:

| Problem | Amazon/Flipkart model | ONDC model |
|---------|----------------------|------------|
| Seller discovery | Only through platform's algorithm | Any buyer app can discover any seller app |
| Commission rates | Amazon India: 3–20%+ depending on category | Variable, lower theoretical floor |
| Platform lock-in | High — data, ratings, infrastructure owned by platform | Data portability, no single platform dependency |
| Small seller access | Significant capital required for ads, Prime eligibility | Theoretically open to any seller with a seller app |

## The Numbers That Make This A Story

By early 2026, ONDC was processing roughly **8–10 million transactions per month** — up from near-zero at launch in 2022.

Food delivery (via Swiggy's integration and other buyer apps) accounts for the largest transaction volume. Grocery and hyperlocal commerce follows. Electronics and high-value categories remain marginal.

For comparison: Amazon India and Flipkart together process hundreds of millions of transactions per month. ONDC's share of total Indian e-commerce is still in the low single digits by value.

## What Actually Changed

Food delivery is the honest success story. ONDC disrupted Zomato and Swiggy's pricing power in specific cities — restaurant commissions from ONDC-native buyer apps ran materially lower than platform rates. Some restaurants in Bengaluru and Delhi reported switching a portion of orders to ONDC channels specifically to capture the commission difference.

But the UX story is harder. Most ONDC buyer apps in their early iterations were slower, less reliable, and had weaker discovery than the incumbents they were meant to replace. When you're competing on "interoperability" against platforms that have spent decades tuning their recommendation engines, the structural advantage doesn't automatically translate to consumer experience.

> India Stack solves the distribution problem. It doesn't solve the trust problem.

The trust problem in e-commerce: returns, dispute resolution, counterfeit goods, and customer service accountability are all harder to enforce in a decentralized protocol where no single entity is responsible for the buyer's experience.

## Where It Works, Where It Doesn't

**Works reasonably well:**
- Hyperlocal food and grocery, where transaction simplicity is high and the ONDC commission advantage is material
- Government procurement use cases, where ONDC integration has been mandated
- Small merchants in Tier 2/3 cities getting first-time digital commerce access

**Hasn't worked yet:**
- Electronics and fashion, where returns/counterfeits require platform-level accountability
- Competing on discovery with Amazon's search quality
- Consumer-facing brand trust — most buyers don't know or care that they're using ONDC

## The Open Question

ONDC's governance is run by the Quality Council of India under the Department for Promotion of Industry and Internal Trade. That's a government structure. The long-term question is whether a protocol built and overseen by a government entity can keep pace with iteration speed against a $1.8 trillion market cap commercial competitor.

UPI worked against private payment players because the incumbents were also fragmented and UPI's trust anchor was the banking system. E-commerce incumbents have deeper network effects, better logistics infrastructure, and clearer accountability structures than ONDC's distributed model has yet matched.

This works in cities for food. It hasn't reached the point where it structurally changes Amazon's or Flipkart's economics — at least not yet.

ONDC vs. Amazon: What Actually Happened When India Built Its Own E-commerce Protocol

// COMMENTS

ON THIS PAGE