ONDC in 2026The Seller Rebellion Amazon Didn't Expect

ONDC — the Open Network for Digital Commerce — is a government-backed initiative launched in 2022 by the Department for Promotion of Industry and Internal Trade (DPIIT) under the Ministry of Commerce. The pitch is straightforward on the surface: build an open, interoperable layer for digital commerce in India, the way UPI did for payments. Strip away the press releases and what you get is a Beckn-protocol-powered network where any buyer-side app can discover any seller-side app without either side needing to belong to the same platform. That single sentence is the entire strategic story. ONDC is not a marketplace. It is not Amazon's competitor in the sense of building a rival storefront. There is no ONDC cart. There is no ONDC checkout page. What ONDC provides is the plumbing — the open specification, the registries, the discovery layer, and the network of network participants (NPs) — that lets a buyer app like Paytm, Magicpin, or Pincode show a product from a seller app hosted by an unknown kirana store in Tirupur or a D2C brand in Andheri. The transaction can complete without either side ever knowing the other's backend. This is the part most sellers misunderstand. ONDC is not where your customers go. ONDC is the layer that makes your customers find you, even when those customers are on apps you have never onboarded to. That is a categorically different growth model from listing on Amazon, and it is the reason the seller economics can be very different too. What ONDC is not: it is not a guarantee of volume. It is not a finished product in the sense of a polished seller dashboard with magic growth. And it is emphatically not a charity from the government — every transaction still has to clear a working business case for the network participants on both sides.

If UPI is what happened when India said "any app should be able to pay any other app," ONDC is the same idea applied to buying and selling. The protocol (Beckn) is the equivalent of the UPI rails. The buyer apps are the equivalent of Google Pay, PhonePe, and Paytm. The seller apps are the equivalent of the merchant-side stack. The user never sees "UPI" on their screen; they see the app. Same with ONDC.

In 2023, "selling on ONDC" was a leap of faith for most SMBs. The buyer apps were thin, the seller app choices were thin, the catalogue rules were a moving target, and discovery was — in the polite framing — uneven. By 2025, the picture started to settle: a handful of seller apps had become serious, the catalogue schema had stabilised, and certain categories — mobility (Namma Yatri being the most-cited example), food, grocery, and B2B — were producing real, repeatable order flow. In 2026, three things are different enough that founders should re-evaluate ONDC, not as a side project, but as a real channel. First, the number of active buyer apps has crossed the point where being on ONDC is roughly equivalent to being on a small set of medium-sized marketplaces at once. Discovery is no longer theoretical. Second, the category coverage has widened into fashion, beauty, electronics accessories, and regional retail, with regional and Tier-2 / Tier-3 commerce growing faster than the headline numbers suggest. Third — and this is the underrated shift — the seller-app ecosystem has matured. A few seller apps now offer storefronts, inventory, and basic analytics that a non-technical founder can operate. That was the missing piece in 2023. None of this means ONDC is suddenly a gold rush. It means the floor has lifted, and the 2026 case for ONDC is no longer "we should explore this someday." It is "we should put a test catalogue live this quarter."

The honest reason a small Indian seller would ever leave — or rather, deprioritise — Amazon or Flipkart is not ideology. It is the take rate, the long tail of platform-controlled costs (storage, advertising, removal orders, returns), and the rigidity of single-platform policy. ONDC's appeal for sellers, expressed in the most founder-friendly framing, is:

• Commission and platform fees that are materially lower than what Amazon or Flipkart charge in the most commonly contested categories.
• No requirement to maintain a single store on a single platform — your catalogue is yours, and you carry it into multiple buyer apps simultaneously.
• A direct relationship with the buyer app, including the ability to negotiate, push back, and reconfigure terms that you cannot negotiate on Amazon.
• A multi-buyer-network world where a single onboarding produces exposure to several buyer-side apps.

We are not going to quote specific commission percentages here because those numbers shift by category, by seller app, and by volume, and the wrong number in a founder's planning sheet is worse than no number. What we will say is that the public material from ONDC and from active seller apps has consistently framed the platform take as a single-digit percentage of transaction value in several retail and F&B categories, with logistics and payment-handling layered on top. Compare that honestly against the all-in cost you are paying on Amazon for the same unit, including advertising, and the delta is rarely a rounding error. The flip side — and founders who only hear the bullish version will be ambushed by this — is that lower take rate does not magically equal more profit. ONDC order volumes are, in absolute terms, smaller than Amazon's for most product categories in 2026. The unit economics are better; the volume is thinner. Whether the trade-off makes sense depends on your category, your SKU depth, and your cost of customer acquisition outside Amazon.

A few honest patterns are emerging from the seller-side of the network in 2026. They are worth naming clearly, because most of the public conversation still skews toward 2023's struggles.

• Hyperlocal F&B is the clearest win. Restaurants and cloud kitchens on ONDC food rails, sold through buyer apps with strong local penetration, are seeing order density that competes with the big food aggregators on net margin, in part because commission is lower and in part because the restaurant's data is theirs to own.
• Mobility remains a flagship — Namma Yatri's open-network model is now the de facto reference point for the whole programme, and it is the proof-of-concept that an open-protocol network can build consumer habit at scale.
• Regional retail and Tier-2 / Tier-3 commerce is the underrated story. Small-town sellers in categories like ethnic wear, regional FMCG, and local services are getting found by buyer apps that they never had to onboard individually. Discovery is doing real work here.
• B2B is the sleeper. ONDC's B2B rails have been quietly growing in categories like industrial supplies, packaging, and HoReCa, where the buyer app is essentially a procurement tool and the seller gets a recurring relationship rather than a one-off marketplace sale.

What is not working, and where founders should be honest with themselves: pure discovery-driven selling of undifferentiated long-tail SKUs in high-competition categories. That is a category where Amazon still wins, and ONDC does not pretend to fix it.

A founder's ONDC plan that ignores the real challenges will fail on a predictable timeline. These are the four that we see most often. Discovery is uneven. Some buyer apps surface your catalogue well. Some don't. There is no single switch you can flip to be everywhere. You will need to monitor which buyer app is generating impressions, clicks, and orders, and adjust. Order volumes are smaller in absolute terms. We said it once and it bears repeating. Better unit economics with thinner volume is not a better business unless you are deliberately trading off Amazon dependence for margin recovery, or using ONDC as a diversification layer. Logistics and returns are still on you. ONDC does not run a logistics network. You either integrate a third-party logistics partner, work with the buyer app's recommended logistics, or run it yourself. Returns management is even more variable — policies differ by buyer app, and "ONDC" itself does not adjudicate. Cataloguing is non-trivial. ONDC-specific cataloguing — buyer-friendly SKU modelling, F&B menu structures, attribute mapping — is its own skill. Seller apps help, but the seller who skimps on catalogue quality will be invisible in discovery.

A founder-friendly onboarding, ordered for a 60-day test:

1. 01
   Pick one product line — not your whole catalogue. Choose the line where you have the best unit economics and the cleanest fulfilment story.
2. 02
   Map it to your nearest ONDC-ready seller app. Do not onboard to three seller apps on day one. Pick the one with the best regional penetration and the most transparent economics.
3. 03
   Build the catalogue properly. Invest in the ONDC-specific attributes. Treat catalogue quality as a one-time cost that pays back in discovery.
4. 04
   Onboard a small test catalogue — typically 20 to 50 SKUs to start.
5. 05
   Run for 60 days against your Amazon storefront. Track three numbers: orders per week, net margin per order, and incremental reach (buyers you would not have reached on Amazon).
6. 06
   Expand or pull back based on those three numbers, not on vibes.

This is the most important paragraph in this entire article. The 60-day, three-number test is the only honest way to know whether ONDC is a real channel for your business or a curiosity.

• Onboarding to every seller app at once. You will spread your catalogue quality thin and learn nothing.
• Treating ONDC as a charity or government programme. It is a network. You are there to make money. The founders who succeed are the ones who run it like a P&L, not a CSR project.
• Ignoring catalogue quality. ONDC discovery rewards well-modelled catalogues. Lazy SKUs get buried.
• Quitting at week three. ONDC ramp is not Amazon ramp. Give it the full 60 days before you judge.
• Betting the company on it. ONDC in 2026 is a serious channel, not a replacement for the rest of your distribution. Treat it as a deliberate diversification, not a leap of faith.

ONDC in 2026 is the open network Amazon and Flipkart could not have built themselves, precisely because their entire business model rests on being a closed network. The opportunity for Indian sellers is real, the unit economics are real, and the network is finally thick enough to bet a small portion of your business on. The mistake is not in joining — it is in joining without a plan, a test, or a clean way to measure whether it is working. Pick one product line, map it to your nearest ONDC-ready seller app, and onboard a small test catalogue to compare conversion economics against your Amazon storefront for 60 days. The honest answer — yes, no, or "yes in this category, no in that one" — is worth more than another quarter of speculation. Immediate Action Step: Within the next 7 days, pick one product line, identify the two ONDC-ready seller apps with the strongest buyer-app presence in your category and region, and shortlist one for a 60-day pilot. Build a clean test catalogue of 20 to 50 SKUs with proper ONDC attributes, ship it live, and commit to a single decision rule: if net margin per order on the ONDC rail is at least 30% better than your Amazon equivalent and the buyer app surfaces real discovery, you scale to your full catalogue. If not, you have lost one quarter of effort, not one year of mojo.