ONDC’s Kirana Digitisation Starts With Buying, Not Selling
DigiDukaan digitises the shopkeeper as a purchaser. The repeat-order rate is the number that tests whether the thesis of kirana digitisation holds.
ONDC and the Department for Promotion of Industry and Internal Trade convened more than forty FMCG chief executives at Vanijya Bhawan this week, with the commerce minister in the chair, to ask them for something unusual: commitment to a procurement network that runs alongside their existing distributors rather than replacing them. The initiative, DigiDukaan, does not help kiranas sell online. It helps them buy. The roundtable, titled Bharat Commerce Chintan Shivir, drew senior leadership from Hindustan Unilever, Coca-Cola India, ITC, P&G, Marico, CavinKare, Pidilite, Britannia, Adani Wilmar, LT Foods, Anmol Industries, Haldiram’s, Emami, Veeba and Nilon’s, alongside logistics providers, technology partners and network participants.
The inversion
Almost every previous attempt at kirana digitisation in India started from the selling side: put the shop on a marketplace, aggregate demand, deliver. That model struggled because a kirana’s advantage is proximity and credit, and an app improves neither. A shopkeeper who already knows their customers does not need help finding them.
DigiDukaan starts from the other direction. A retailer orders stock across multiple brands and distributors through one interface, with visibility into pricing, trade schemes and availability. The claim is that this addresses a daily friction. A shopkeeper currently reconciles several distributor relationships, each with its own ordering method, scheme structure and delivery schedule, largely on paper and memory. Consolidating that is a smaller promise than reinventing retail and a more useful one.
Whether the inversion works is an empirical question, and the release supplies one figure that tests it.
Why the network model matters here
ONDC’s premise is that a protocol rather than a platform should mediate commerce, so a buyer on one application can transact with a seller on another without either joining the same marketplace. Applied to procurement, that means a retailer’s ordering application, a brand’s distributor system and a logistics provider can interoperate without a bilateral integration for each pair. A shopkeeper buying from eight distributors across two dozen brands would otherwise need sixteen separate integrations. The open-network model is the only one in which the marginal brand is cheap to add, and cheap marginal addition is what kirana digitisation at national scale requires.
The numbers that matter
DigiDukaan reports 21,000 retailers connected to more than 450 brands across Hyderabad, Jaipur and Mumbai as of September 2026, up from 12,000 in the previous quarter. In Hyderabad, where more than 16,000 retailers are onboarded, monthly repeat ordering runs at 73 to 78 per cent.
That repeat rate is the most important figure in the document. Acquisition numbers in Indian retail pilots are easy to produce with incentives, and a shopkeeper who signs up for a discount tells you nothing about whether the product works. A repeat rate near three quarters means retailers returned the following month. If it holds as the network expands beyond a single dense city, it is evidence of habit rather than trial, and habit is the only thing that makes a procurement network defensible.
The second figure is softer. An ONDC retailer study found 40 per cent of surveyed retailers reported lower inventory holdings for products procured through the platform. That is a working-capital claim, and it is the one a shopkeeper would care about most, since money tied up in stock is money not available for anything else. But the release gives no sample size, no methodology and no comparison group, so the figure describes a survey rather than demonstrating an effect.
What the roundtable asked for
The five priorities read like a list of what the network still lacks rather than what it has achieved, which makes them more informative than the achievements. A common digital identity for kiranas, so a retailer does not re-register with every brand, distributor and service provider. Shared logistics and micro-warehousing, aimed at next-day replenishment for outlets outside existing delivery routes. Formal credit linked to procurement and transaction history, reducing dependence on informal distributor credit. Plug-and-play distributor management system connectors and standardised catalogues, to lower the cost of joining. And a review of how sales teams, distributors and channel partners are incentivised.
The fifth is the candid one. A distributor whose sales force is measured on volume through its own order book has little reason to encourage retailers onto a network that lets them buy from a competitor. Naming incentive realignment as a priority concedes that the channel has a reason to resist, which is more honest than most platform announcements manage.
The incentive problem
This is where comparable initiatives have struggled. A business-to-business procurement network works only if distributors see more reach than they lose control of, and the release’s framing reflects that tension: strengthen existing distribution relationships rather than replace them, shared infrastructure rather than disintermediation.
The logistics priority points the same way. Next-day replenishment for outlets outside existing routes is a service a single distributor cannot easily provide and would value from a shared layer. If DigiDukaan sells that, it earns a place in the channel. If it only moves orders, it competes with the distributor’s own order book, and the distributor has no reason to cooperate.
Who is actually live
The release distinguishes participation from integration, and the distinction deserves care. Hindustan Unilever, Marico, Colgate-Palmolive, Emami and Amul are described as participating through their distribution networks. ITC, Tata Consumer Products and Britannia sit at various stages of integration or discussion. The first group has a live path to the network. The second has intent.
On the technology side, Paytm for Business, Qwipo and Mulltiply.ai are live, with Bizom, Botree, QuIDPe and ENS progressing towards integration. Because most FMCG ordering in India runs through a distributor management system, the connector question is the real bottleneck. A brand cannot join at scale until its system can talk to the network, and the release lists more providers in progress than live. That is the single most useful sentence in the document for anyone assessing the timeline.
How retailers actually get recruited
The National Traders Welfare Board receives one sentence in the release and deserves more. Mobilising local trade associations is how retailer recruitment works in Indian general trade, where a shopkeeper adopts what the association in their market adopts. A network that relies on digital acquisition alone reaches the organised end of the market; one that works through trade bodies reaches the rest. That mechanism decides whether kirana digitisation scales beyond the top cities, and it is the part of the programme with no published metrics.
The 1.4 crore framing
The release describes DigiDukaan as digitising procurement for India’s 1.4 crore kiranas. That figure is the size of the national retailer base, not the addressable market for a procurement network, and the gap between the two is wide. A shop that buys one category from one distributor has little reason to join a multi-brand ordering platform, however good the interface.
The cities chosen show where the model is being tested: Hyderabad, Jaipur and Mumbai, with the next phase concentrated on North Mumbai. These are dense, distributor-rich markets where multi-brand ordering saves a shopkeeper real time. Whether the approach works in a smaller town with two distributors is the question the current numbers cannot answer, and the answer determines whether kirana digitisation reaches the country or the top forty cities.
What to watch
Three markers. The first is whether the repeat rate holds in Mumbai, since the Hyderabad figure comes from a single dense market where the network runs deepest and where a shopkeeper has the most brands to choose between. The second is distributor management system integration: the number of brands moving from discussion to live tells you more than the number attending a roundtable. The third is whether the credit priority becomes a product, because formal credit tied to procurement history is the benefit that would keep a shopkeeper on the network after the novelty and the introductory schemes fade.
The wider portfolio
The wider ONDC portfolio gives a sense of what the network already carries: more than five lakh metro and bus journeys a day through forty consumer applications, logistics infrastructure used by 80,000 merchants and thirty small fleet operators, and a thousand tourism experiences including 170 ASI monuments and 29 museums. Procurement is the least visible of these and the one with the largest transaction value per participant, which is why the buy-side bet is the more interesting one.
What it means
ONDC has processed more than 500 million transactions across its network, with retail as its core focus. DigiDukaan is the first initiative in that portfolio aimed squarely at the buy side of Indian retail, and it carries the clearest commercial logic of the set: a shopkeeper’s purchasing is a larger and more predictable transaction than their selling. The commerce minister’s framing was that adoption follows when someone builds open infrastructure around the small retailer’s needs. Three things would have to hold for kirana digitisation to travel.
Distributor management systems would have to connect fast enough that a brand can join without custom work. Distributors would have to find the shared logistics layer worth more than the order book they give up. And a shopkeeper would have to keep ordering after the introductory schemes end, which is what the repeat rate measures. The release supplies evidence on the third and a to-do list on the first two.