India is about to crack digital commerce wide open

UPI solves for a relatively simple problem, payments. It is the same service over and over, with a relatively limited number of stakeholders involved. There are the banks, the aggregators (Paytm, PhonePe, Google Pay and others, but still a very concentrated part of the chain) and the public agency NPCI who owns the digital platform. Simple problem, simple and very elegant solution.

ONDC’s scope is at another level of complexity. An endless number of products and services (food, transportation, finance, retail, etc) with thousands and thousands of sellers and providers of related services (logistics, finance, customer service, etc), all connected through API links. For instance, the percentage of payment failures on the UPI network stands at around 2% while the percentage of e-commerce purchases that are returned (and thus require a lot of follow-up actions) are estimated at 25%-40%.

Two obvious challenges emerge from this increased complexity. First, it requires much more coordination, which the platform needs to provide. For instance, who solves a return? A reimbursement? Who takes responsibility for a seamless customer experience? A retailer like Amazon spends a lot of money to ensure that all those processes work close to perfect. Theoretically, one of the key roles of an ecosystem orchestrator (platform aggregator) is to ensure a satisfactory interaction. That requires great matching (the buyer finds the best seller for any specific need) but also great interaction between the two (delivery, conflicts, queries, payments, etc.) ONDC, for now at least, seems to be solving only for great matching, while interaction quality is left to the individual participants. That is always a risk.

Second, ONDC’s complexity makes cost efficiency challenging. Can a network with so many small companies operating at small scale have the same unit costs than a large marketplace like Amazon? For instance, UPI is already looking for some revenue streams, after starting to feel the limits of offering something completely free. And in that case, we are talking about companies that are very large and well-funded (Paytm, PhonePe, banks, etc). ONDC, at scale, would likely have to be subsidized at a whole other level to sustain the low prices in the above graph.

Another potential difference between ONDC and UPI is the “ceiling” of their respective markets. There were just a few million users of digital payments in India before UPI came in and opened access to anyone with a phone. In a matter of months, it was clear that this was an attractive option to hundreds of millions of Indians (260M+ and growing fast). Does ONDC have that kind of potential in terms of number of users? Right now, as noted above, there are around 185 million e-commerce users. But it is a heavily skewed distribution, as 50% of the online commerce spending is done by only 45 million of those users. If ONDC brings in more sellers and perhaps slightly lower prices, will that 45 million number explode into hundreds of millions?

On one hand, the number of use cases and potential combinations that a network like ONDC offers should foster a lot more online purchasing of products and services. Expanding B2C marketplaces, increasing B2B transactions (think for example how farmers could access many more vendors for equipment and other key inputs), fueling P2P commerce (e.g., used goods, apartment/room rentals, etc.), or boosting D2C commerce (any household becomes a potential clothing or home cooking or education brand, to name just a few) are some of those use cases.

On the other hand, if there are only 45 million “regular” e-commerce users in India right now, is that a problem of not enough supply? UPI enabled digital payments for consumers who were not able to make such payments before. But ONDC seems to be solving mostly for supplier access, not consumer access, which is already high (see the 185 million figure of total transactors). The gap between 185M total users and 45M regular users and the fact that the current online offer is already very attractive (e.g., deep online discounts, waving of delivery costs) raises the question of whether the problem is really the willingness of consumers (especially for certain types of goods) to buy online. If that were the case, then the potential market size for ONDC would likely be smaller than some of the optimistic estimates.

In sum, ONDC is a fascinating experiment that is worth every bit of attention in the coming months and years. Many elements (5G adoption, AI, combination with other digital networks, country growth, adoption abroad, etc) might multiply its effect and make it another world-class innovation. But in any case, ONDC is different enough from UPI to deserve its own analysis.