What ONDC Promises
Open Network for Digital Commerce (ONDC) is the Indian government's ambitious project to create an open, interoperable e-commerce network. The pitch: any buyer app can discover products from any seller app, breaking the monopoly of Amazon and Flipkart. It's UPI for e-commerce. On paper, it's brilliant. In practice, we've spent six months integrating two clients (one buyer-side app, one seller-side platform) and the experience has been... mixed.
Let us share the technical reality, without the PR spin.
The Protocol Is Complex but Well-Documented
ONDC uses the Beckn Protocol, which defines APIs for: search (buyer app queries the network), on_search (seller apps respond with catalogs), select/init/confirm (order flow), and post-order flows (tracking, cancellation, returns). The documentation is thorough — probably the most complete API documentation we've seen from an Indian government initiative. The Beckn protocol spec is on GitHub, and the ONDC-specific adaptation is documented separately.
That said, the spec is large. Implementing a complete buyer-side or seller-side integration involves about 25 API endpoints, each with a specific schema, signature verification (every request is signed with the participant's private key), and callback handling. Our initial estimate was 4 weeks for integration. Actual time: 10 weeks. The delta was mostly in edge case handling and certification.
The Certification Process Is the Real Bottleneck
Before going live on ONDC, your application must pass certification by the ONDC team. This involves: technical compliance testing (do your APIs conform to the spec?), functional testing (does the order flow work end-to-end?), and security review. The certification process took us 5 weeks for the first client. The feedback cycle is slow — you submit, wait 5-7 business days for feedback, fix issues, resubmit. Some feedback was about genuine technical issues. Some felt like subjective interpretations of the spec where our implementation was valid but not how the reviewer expected it.
Our advice: join the ONDC Slack workspace (yes, they have one) and build relationships with the technical team. Direct communication reduces the feedback cycle significantly. Also, run the ONDC reference implementation alongside yours and compare API responses — if your response differs from the reference implementation, the certification team will likely flag it, even if the spec technically allows your format.
Catalog Management Is the Hardest Part
For seller-side integration, maintaining an ONDC-compliant catalog is the biggest ongoing challenge. ONDC requires: structured product attributes in a specific format, category-specific required fields (electronics need brand, model, and warranty info; food items need FSSAI license, expiry date, and nutritional info), pricing that includes MRP, selling price, and all applicable taxes broken out, and real-time inventory updates (you can't sell something that's out of stock on ONDC without risking penalties).
Our client (a mid-size retailer with 5,000 SKUs) had product data in an ERP system with inconsistent formatting. We built a catalog sync pipeline: ERP → data cleaning/enrichment → ONDC-formatted catalog → published to the network. The data cleaning step is surprisingly labor-intensive. About 30% of products needed manual attribute enrichment because the ERP data was incomplete. For new merchants joining ONDC, budget significant time for catalog preparation.
The Logistics Integration Gap
ONDC supports three fulfillment models: seller-fulfilled, buyer-app-fulfilled, and third-party logistics. In theory, logistics providers are part of the network and fulfillment is handled seamlessly. In practice, the logistics network participant ecosystem is still thin. During our integration period, there were only 4-5 active logistics providers on ONDC, and their coverage was limited to metro cities.
For our client who operates in Tier 2 cities, we ended up using seller-fulfilled delivery with their existing logistics partners, and exposing the tracking information through ONDC's tracking APIs. This works but loses the "network effect" benefit that ONDC promises for logistics.
Is It Worth It?
For sellers: if you're already selling online and want another channel with potentially lower commission rates (ONDC doesn't charge platform commission like Amazon's 15-25%), the integration cost is justified. Budget 10-12 weeks for a complete integration. For buyer-side apps: the catalog diversity on ONDC is still limited compared to established marketplaces. It's growing, but slowly. We'd recommend ONDC integration as a complementary channel, not a primary one. The technology works. The network effects aren't there yet. But when they arrive — and we believe they will, given the government's push — being already integrated will be an advantage.