116,000 ONDC Sellers and Almost Nobody's Actually Selling

2026-07-05 — sales strategy small business India

I walked through a marketplace in Bengaluru last week and talked to a shop owner who'd registered on ONDC six months ago. He'd never received a single order.

The headline numbers look good. As of December 2025, India's government confirmed more than 116,000 retail sellers live on ONDC spread across 630 cities and towns. If you run a kirana in Coimbatore or make textiles in Jaipur, the network is theoretically yours to use. No Amazon taking 25 percent. No Flipkart taking 20 percent on fashion jewelry. Just you and a digital infrastructure built by the government, not by venture capital.

And for 63 million MSMEs in India?

That matters.

But here's the thing that matters more. When Anitva analyzed the numbers in April 2026, they found approximately 85 percent of those registered sellers weren't active. Not struggling. Not selling at a loss. Registered and then abandoned.

The Math Does Work, at Least on Paper

The cost case is real, I'm not being clever. Amazon's referral fees range from 5 to 25 percent depending on category. Flipkart runs something similar—3 percent on mobiles, 25 percent on jewelry. ONDC charges no central commission. The costs still exist technically, distributed through seller network participants, but most small sellers end up paying considerably less. Actually, let me correct that—they end up paying considerably less if they have transactions at all. If you're selling handmade textiles at a 20 percent margin, the difference between a 15 percent platform cut and something closer to 3 percent is whether your business survives or not.

The GST situation helps too, or helped, something like that. From October 2023, sellers with turnovers below ₹40 lakh in most states could register on ONDC using an enrollment number instead of a full GSTIN. One less bureaucratic obstacle. The Ministry of MSME's TEAM scheme added digital literacy support, cataloguing help, photo assistance. Actual hand-holding instead of just a portal. That's more or less how you get someone's grandmother's pickle business online.

All of this is real.

The Real Problem Isn't Registration

Amazon and Flipkart have fifteen years of consumer habit. They have logistics networks. They have customer trust embedded in millions of phones. ONDC's architecture means a buyer on Meesho or Magicpin might theoretically reach a seller on the network, but the discovery layer is thin and inconsistent. Thin might be generous. A seller who registers in December and sees zero orders by February stops checking the app. That's how you get 85 percent inactivity. That's how you get the shop owner in Bengaluru I mentioned, the one who hasn't received a single order in six months.

The infrastructure exists.

The cost structure exists.

Whether enough Indian consumers actually buy through ONDC-connected apps to make selling there worthwhile is still open. We're halfway through 2026 and the network has 630 cities covered. Getting 116,000 sellers to their first hundred orders is a different problem than infrastructure. No government scheme has solved that yet. Most consultants, including us sometimes, skip over that part and talk about the vision instead.

The question isn't whether ONDC is real. It's whether a seller in Agra can actually feed their family using it.