पाठशाला Pathshala · वृद्धि Vṛddhi, Growth · Lesson 17 · Build

ONDC and the open network: a channel or a distraction

The Open Network for Digital Commerce lets any buyer app sell any seller’s goods. Read the volumes carefully, price what it costs to join, and decide whether it is a channel for your category or a distraction from one.

Pathshala, The Founder Library · 11 October 2026 · 7 min read

Chinese fishing nets stand over the water at Kochi while fishermen work on the shore. — Photograph: Ravi Kant · Pexels

A Coimbatore snacks brand is told by three different people in one week that it must get on ONDC. One is a seller app’s sales team, one is a government seminar and one is a competitor who says orders are pouring in. Nobody can tell the founder how many orders a brand like hers would get in a month, or what each would cost. That is the question this lesson answers.

The Open Network for Digital Commerce is the most ambitious attempt anywhere to make e-commerce work like UPI: open rails that any buyer app and any seller can use, instead of a few platforms that own both sides. For a founder it is a distribution question, and it deserves the same treatment as any channel: what it is, what volumes it carries in your category, what it costs to join, and how to test it without spending a quarter on it. Every figure below is as reported by the source named, and the network changes fast; the figures were checked in October 2026.

What the open network is

ONDC is not a marketplace. It is a set of shared specifications, an initiative of the Department for Promotion of Industry and Internal Trade, through which separate applications talk to each other. Kalaari Capital’s ONDC primer sets out the roles. Buyer apps are the consumer-facing applications where people search and order. Seller apps hold sellers’ catalogues and inventory and handle fulfilment. Gateways route searches so that every seller app has a fair chance of being found, and registries list who is on the network. Logistics providers pick up and deliver, and a seller can use one on the network or bring its own.

A brand does not join ONDC directly. It signs up with a seller app, gives it its catalogue, prices, stock and GST and bank details, and the seller app makes those products discoverable in every buyer app that serves the category. The promise to the seller is reach without exclusivity: one listing, many storefronts, and terms set by a common framework rather than by a single platform. The [India Stack lesson](/library/india-stack-as-market-map) places it beside UPI and the account aggregators as public rails a company can build on.

The rules of discovery are part of the design. The government’s December 2025 statement to the Lok Sabha described seller apps making their full catalogues discoverable to every buyer app, and buyer apps disclosing the main parameters they use to rank search results, so that a seller can see why it appears where it does and improve its position. That is a real difference from a single marketplace, whose ranking is its own business. It also means the work of being found is the seller’s: complete catalogue data, accurate stock, honest photographs and prices, and fast acceptance of orders are what a ranking can see. A seller who lists badly on the open network is invisible on every buyer app at once.

The volumes, read carefully

The government’s own trade portal reported 12 million orders in July 2024, with a daily peak of about 430,000. The split matters more than the total: 4.4 million were mobility, mostly rides; of the 7.6 million others, food and beverages were 1.7 million, groceries 1.4 million and fashion 0.6 million, with the rest spread across logistics, beauty, home and kitchen and retail vouchers. About 6.3 lakh sellers and service providers were on the network then. October 2024 reached 12.96 million orders, MediaNama reported.

Then the retail line turned. Inc42’s analysis in April 2025 reported retail orders peaking at 6.5 million a month in October 2024 and falling to 4.6 million in February 2025, the lowest in ten months, with groceries about 1.8 million of that. Over the same months network incentives per participant were cut from ₹3 crore in July 2024 to ₹30 lakh in December, and buyer apps reduced the discounts that had driven grocery and food orders. Logistics went the other way, from 35,000 orders in April 2024 to 2 million in February 2025, and mobility held up. The network crossed 200 million cumulative transactions by March 2025. In December 2025 the Minister of State for Commerce told the Lok Sabha that more than 1.16 lakh retail sellers were live across more than 630 cities and towns. Note the definitions: retail sellers live is a narrower count than sellers and service providers onboarded, and both are different from sellers receiving orders.

What it costs to be on it

The fees are lower than a marketplace’s, and they are not zero. From January 2025 the network charges sellers ₹1.5 on each successful transaction above ₹250, with cancelled orders exempt, according to the MediaNama report. The buyer app takes a fee for bringing the order, which the protocol calls the buyer finder fee; Kalaari’s worked example on a ₹500 order has the buyer app taking ₹15, 3 per cent, the seller app ₹10, 2 per cent, and the seller keeping ₹475, with delivery charged separately. It compares that with marketplace commissions of 15 to 30 per cent depending on the marketing and fulfilment included, and Inc42 reported sellers paying ONDC buyer apps about 5 per cent against 30 to 40 per cent on the large food platforms.

The cost that matters more is time. Going live through a billing, POS or inventory provider that already acts as a seller app can take days. Building a direct integration means keeping catalogue, prices and stock in sync with every buyer app, handling orders, cancellations and returns within the network’s timelines and answering complaints that arrive through it, and that is a small team. Delivery still has to be arranged and paid for on every order. The incentive cuts are the reminder that a channel whose volumes are paid for by someone else can change when that someone changes their mind.

Where it has worked and where it has not

The evidence points one way. Services the seller runs itself, rides and deliveries, have used the network well because many buyer apps can send orders to the same fleet; Kalaari’s primer reports Namma Yatri, the Bengaluru ride app built on the protocol, completing about 6 lakh trips in six months. Local food and grocery grew fast on discounts and shrank when the discounts went, while, by Inc42’s account, quick-commerce apps outside the network took the same customers. Packaged and fashion brands shipping nationally have had a cheaper listing than a marketplace but far fewer buyers searching. The walk-through below turns that evidence into a decision for your category.

A street vendor stands beside his snack stall in a busy Indian market. — Local food sellers joined the network in large numbers. Whether the orders stayed depended on whether buyers came without a discount. Photograph: Anubrata Saha · Pexels

A channel is the place your buyers already search. ONDC is a channel where they do, and a press release where they do not.

A ninety-day test

If the tree says test, run it like any channel experiment from the [first channel lesson](/library/finding-first-channel-that-works). Go live through a seller app you already use, with your twenty best-selling products, real photographs and prices no higher than on your own site. Write three targets before launch: orders a week by day ninety, contribution per order after every fee and delivery, and the share of buyers who order a second time within sixty days. Check which buyer apps the orders come from, because a channel that is really one buyer app running a promotion is that buyer app’s channel. At day ninety, keep the channel if it met two of the three targets, and stop if it met none; do not extend a test because the network has announced something new.

Two Indian specifics make the test fair. Use the Trade Enablement and Marketing scheme of the MSME ministry if you qualify, which the December 2025 statement described as supporting small sellers with cataloguing and onboarding on the network. And price delivery honestly: an order that is profitable when the customer collects it is often loss-making when a courier carries it across a city.

The quarterly channel line

Whether you are on the network or not, give it one line in the quarterly channel review. If live: orders, contribution per order, repeat rate and orders by buyer app, compared with the same figures for your marketplace and your own site. If not live: the latest reported volume for your category and the fee terms, with the date. Rerun the tree each quarter when either moves; a channel that was a distraction in 2025 can become a channel later if buyer apps start to bring orders without subsidy. Then make the decision in one sentence and move on to the channels that are working.

Figures are as reported by the sources below, which are dated; ONDC’s fees, incentives and volumes change often. Confirm current terms with your seller app before deciding. Nothing here is investment advice.