Published on Tech in Asia.


As internet access improves and more people get smartphones, ecommerce startups are placing massive bets on India. Global giant Amazon competes against homegrown hopeful Flipkart, which is tailed closely by Snapdeal. There are many others, including Shopclues, fashion site Myntra, and furniture startup Pepperfry. Their joint challenge? Proving to hundreds of millions of shoppers that buying online is more convenient, cheaper, and provides more options than buying things in person.

The problem is, India is unique, with challenges that countries like the United States don’t present. Visit any city and you’ll see hundreds of small shops dotting every corner. They come in all shapes and sizes and sell everything one can imagine – medicine, toothpaste, gifts, used computers. All take cash, some take cards, and a few have even started adopting e-wallets. What’s most important is that they interact with their customers enough to know them well and have conventionally bent over backwards to serve them – like delivering goods to doorsteps or installing electricals within houses.

That’s where the enigmatic and oft-quoted India-specific challenge appears. As ecommerce companies pump money into winning over customers that aren’t top-tier and easy targets, it’s clear that a major chunk of their competition comes from offline stores.

Right in the midst of these poles is Dunzo, a small startup that works out of a converted home in Indiranagar, Bangalore. At any hour of the day, you’re guaranteed to find at least one of its forty employees typing away at a keyboard, ready to complete a task.

I first came across Dunzo when a friend needed chairs for her balcony. We looked online but didn’t want to wait days for the order to be delivered, or bet on shaky logistics that might mean that the package would come at odd hours – or worse, would never come. We found local places on listings site Justdial but it was hard to tell which were open, or which weren’t located in shady parts of town, or which would fit with our busy schedules. That’s when someone suggested Dunzo – I downloaded it on a Sunday morning, chatted with a representative for 5 minutes, and had the chairs delivered by that afternoon.

The chairs came from a store closeby – Dunzo charged me just thirty rupees (half a dollar) on top of the retail price to have it delivered, because it was less than three kilometers away. The representative I chatted with sent me photos of chair designs I might be interested in. I insisted on the cheapest possible option. I chose to pay cash-on-delivery, but could have gone for net banking or credit and debit payments.

Impressed, I had a friend connect me with the team and set up a meeting with co-founder Kabeer Biswas. “We give you superpowers,” he said to me, staunchly.

Demand first

Kabeer Biswas is a walking, talking encyclopedia of Bangalore. He knows everything there is to know about the landlocked southern city, and is reasonably embarrassed when I bring up places he doesn’t know. His co-founders are Ankur Aggarwal, Dalvir Suri, and Mukund Jha.

Photo credit: Chris Schmich.

Dunzo first started out as a WhatsApp group, he reminisces, an experiment for Kabeer and his friends after quitting their jobs. It was as beta as beta testing gets and limited to a few customers. The demand was enough that Kabeer and others began to get stressed out – which meant that it was high time that it was turned into an app.

The first and most important takeaway that he provides me with is that their initial strategy was to bring customers to merchants. This, he explains, has made all the difference. “Instead of going to merchants and telling them – ‘hey, we have this great idea, sign up with our tech,’ we just went to them saying – ‘hey, we have customers that want to buy from you. We’ll do the payments and the fulfillment’.”

That already makes them different from other startups that have targeted India’s merchants. Take, for example, Amazon, which unveiled Kirana Now in May, an initiative to let local small groceries upload their inventories online and sell them via the site. The initiative has been largely unsuccessful and, if rumors are to be believed, the service has been quietly shut down.

Tech is not to be treated as a blessing, Kabeer explains – especially for those that are already operating well without it.

“What merchants are good at is running their own shops,” he says. “You’re not there to help them with that. We’re just here to add demand.”

He makes an example of an old computer shop, once a staple of Bangalore’s landscape, that increased its revenue by INR7.5 million (US$11,000) after starting to work with Dunzo. That was forty percent of its total revenue for the month.

“You can’t start by tying up with 1,000 merchants,” Kabeer says. “Pick 250 good ones. We should have one electronics guy in an area of Bangalore instead of toying between four or five. Get great merchants on board and you can ensure better traffic and behavior.”

Easy, breezy logistics and payments

Anyone who has been paying attention to Indian ecommerce knows that two of its most difficult bumps are handling payments and managing the logistics of shipping goods. Most companies burn cash and find it hard to become profitable because they waste so much of their resources on these aspects.

Kabeer dismisses those two points when I ask him about them. “We have two goals,” he says. “Goal one: don’t lose any money on logistics. Goal two: make it so that the cost of operations in a city is paid for by the customers.”

Dunzo does not use a logistics company. All delivery people are independent contractors. Kabeer says that Dunzo has about 120 signed on, and each gets paid per task – usually about INR80 (US$1.20) an hour. “They’re generally college students trying to make an extra buck,” he adds.

Once, a customer wanted to order 15GB of RAM. It would take a few days through Amazon, so instead, he enlisted Dunzo to go pick it up from India’s electronics road

But it’s the internal algorithm of Dunzo that powers the multifaceted app.

Let’s say you’re craving pani puri (an Indian street food) from your favorite stand. A Dunzo delivery person can get it done. You can pay him in cash or card, but he will probably buy from the stand owner in cash. But, let’s say you want pigeon spikes installed on your balcony so the pesky birds stop defecating all over it. Then, Dunzo will get in touch with the merchant, who will come to your house, install the spikes, and take payments from you if you’ve selected cash, or get paid by Dunzo if you’ve done it over card.

“It’s faster than Amazon Prime,” Kabeer says cheekily.

At the final point, all payments are transacted on the Dunzo system through card. “It’s an insane number of microtransactions,” he says. “Strangely ranging all the way from INR30 to INR50,000.”

Those variables make me uncomfortable, because it seems as if it would be difficult to scale in a sturdy way. I ask Kabeer about this, but he doesn’t indulge those questions. “Right now, we’re only focused on making Bangalore profitable,” he says. “Nothing more.”

Kabeer is surprisingly candid about those metrics – something rare among India’s internet startups, and a tactic he enjoys flaunting.

The algorithm for profitability has already been figured out, he explains. For a trip that’s shorter than three kilometers, Dunzo charges INR30 (US$0.45). It charges INR60 (US$0.90) for anything over that. Every time it charges INR60, it makes a profit – after paying the delivery person a fee.

Profitability in Bangalore happens at 4,500 to 5,000 transactions a day – currently, Kabeer places it at about 700 to 1,000.

Most importantly, Kabeer stresses that he’s more curious about how Dunzo can stay light on its feet as it expands.

“Pricing is an emotional problem,” he quips. “But pricing at scale is a different algorithm altogether. You ask questions like – what about peak pricing? How can I update the model to fit with growth?”


Chat, as China has so popularly shown with its all-in-one chat app WeChat, has the potential to be immensely disruptive – but only if done right.

It’s malleable because chat apps can be easily updated. They integrate photo and video into a user’s stream of consciousness well, and it’s something that most users in India are already used to. “If WhatsApp actually opens up its API, we will totally partner with them,” Kabeer laughs. “Nobody else, though.”

Dunzo’s app also integrates artificial intelligence that learns from text-based commands. Kabeer claims that 60 percent of transactions are completed via artificial intelligence.

Once a task has been completed, it can be automated. That means I can press “repeat” any time on the task with the chairs, and have them re-delivered to my house. Dunzo will also input those commands into its system to enhance its artificial intelligence technology. If someone orders chairs in language that’s somewhat close to what I used, its AI bot will immediately take care of it.

It’s also not as simple as it looks.

If someone’s spelled “hemming” incorrectly, or botched the name of popular Vietnamese restaurant Phobidden Fruit – Kabeer says one user once spelled it as “phobedden fruit” – it needs to have the breadth of knowledge necessary to understand it.

Even more difficult is understanding context. There are queries like “roll pressing” a sari. Most of the time a roll indicates a type of food – so if the AI interpreted that wrong, it could be potentially disastrous.

Kabeer explains that the Dunzo app will look vastly different in a few months. It wants to introduce features like group chat, where you can collaborate with friends to figure out how to get things done. Another – I couldn’t tell whether this was a joke – might allow for video while a Dunzo employee is moving around.

Will it work?

Photo credit: Juan Antonio F. Segal.

Dunzo is taking on an immense challenge. Scaling the model is difficult. Imagine moving into new cities – will it be so easy for the team to intimately understand the inner workings of, say, Mumbai before serving customers there?

Companies like Flipkart and Amazon also have a lot more money – Dunzo has raised just US$650,000 in seed money. There’s a major chance that larger companies will eventually pump enough cash into India to get people into ecommerce and away from small and medium businesses – before Dunzo has a chance to prove that its technology can be built out in a very India-specific way.

They’ve also got a tighter control over inventory. What if Dunzo’s usual jewelry store doesn’t have the right type of earrings in stock? Will it run around until it finds it?

Its competitors are also not just those that are in ecommerce. Since Dunzo caters to both goods and services, it fights off other local incumbents, like home services startup Housejoy. Globally, other startups also fit the fulfillment criterion. These might pose a threat to Dunzo, which eventually wants to work across the world. San Francisco-based Magic, for example, lets you hire a human assistant for US$249 a week.

“Yes,” answers Kabeer. “These are problems that we will always solve.”

It’s perhaps his attitude that makes Dunzo’s pitch so believable. “Money doesn’t matter because you can’t outspend anyone in this business,” says Kabeer. “If people aren’t interested in what you’re doing, no amount of money can fix that.”

There are a few other important reasons why it might be successful. One – India’s storage-conscious, low-cost smartphone owners are more likely to download an app that has many options over one that only offers a few.

Dunzo also foregoes regulations that the Indian government has otherwise placed on ecommerce, like those that leave selling pharmaceuticals online in a gray area. Instead, Dunzo picks up medicine directly from authorized sellers.

As of now, 40 percent of customers are repeat users, Kabeer claims. He also says that Dunzo has never spent money on marketing – just word of mouth through really great experiences. It’ll never advertise discounts, he says – the product is addictive enough to keep customers around.

Dunzo’s investors also found it after using the app, he claims. One is Rajan Anandan, Google India and Southeast Asia head. Another is Aspada Ventures, which conventionally invests in startups that seek to merge social impact with profitability. There’s also mainstream investor Blume Ventures.

There’s a last interesting example that Kabeer cited that I think showcases Dunzo’s potential best. Once, a customer wanted to order 15GB of RAM. It would take a few days through Amazon, so instead, he enlisted Dunzo to go pick it up from India’s electronics road. “It may have cost a little bit more, but he got it faster and from a closer source,” reminisces Kabeer. That strength is where Dunzo has a chance to succeed.

I ask – are there things Dunzo won’t do? “Well yes, illegal things,” says Kabeer. “We also won’t make custom cakes,” he adds mysteriously.