Start manually and unscalably to prove demand before building infrastructure, but remember that becoming a beloved household verb never guarantees a working business if the underlying unit economics never close.
Dunzo
Dunzo began in 2014 as a WhatsApp group in Bengaluru where founder Kabeer Biswas ran errands for friends, and grew into India's pioneering hyperlocal concierge, the first Indian startup to take direct investment from Google. Its stories run in two directions at once: a manual, unscalable start that became a beloved category-defining brand ("just Dunzo it" turned into a verb), and a late pivot into cash-burning quick commerce that ended in a 2025 shutdown. Brand love and category leadership, it turns out, are not the same thing as a working business.
Cold Start: from a WhatsApp group to a delivery network
The problem. A "deliver anything, run any errand" service is a two-sided cold start: with no couriers there is no reason to ask, and with no requests there is no work for couriers. Building an app, a fleet, and demand simultaneously in 2014 Bengaluru would have required capital and proof Dunzo did not yet have.
The approach. Biswas skipped the product entirely and started manual, running errands for friends through a WhatsApp group where users simply typed what they needed. Dunzo did the tasks by hand, learned the real demand (groceries, forgotten chargers, pharmacy runs, document pickups), and only built the consumer app in April 2017 once patterns and loyalty were established.
How it solved it. The manual loop compounded: roughly 70 tasks a day by June 2015, then over 500 orders a day by the end of 2016, all through WhatsApp before any app existed. That organic, proven traction is what attracted a $650k round in 2016 and, in December 2017, a roughly $12 million investment from Google, its first direct bet on an Indian startup.
Positioning: "anything, anywhere" and a brand that became a verb
The problem. Dunzo was not a single-category app like a food or grocery delivery service. It needed to occupy a broad, memorable position ("we get anything done for you") without becoming a vague, forgettable catch-all in users' minds.
The approach. Dunzo positioned itself as the concierge for a whole city, the "logistics layer of every city," able to pick up and deliver practically anything 24/7, and wrapped it in the tagline "Just Dunzo It" (the name itself is slang for "done"). It leaned into everyday, relatable use cases: the wallet left at home, laundry, ingredients for dinner.
How it solved it. The positioning stuck culturally: "Dunzo it" became a verb Bengaluru residents used for any quick errand, giving the company category ownership across nine major Indian cities. As The Ken later put it, "Dunzo fans made it a verb," a brand equity so strong it outlived the company, though, as its finances showed, love for a verb did not translate into profit per order.
Pivots: chasing 19-minute quick commerce
The problem. By 2021, well-funded rivals (Swiggy Instamart, Zepto, Blinkit) were racing to build instant grocery delivery, and Dunzo's flexible concierge model looked slow and low-margin against 10-to-20-minute promises. Standing still risked ceding the fastest-growing delivery category to competitors.
The approach. In August 2021 Dunzo pivoted hard into quick commerce, launching Dunzo Daily with a 19-minute grocery promise and rapidly building out a network of roughly 130 dark stores. In January 2022 it raised $200 million from Reliance Retail for a 25.8% stake to fund the expansion.
How it solved it. It did not: the pivot torched the unit economics that the concierge model had, at least, kept lean. Dunzo was reportedly burning over ₹230 per order and spending upwards of ₹100 crore a month; losses hit roughly ₹464 crore in FY22 and about ₹1,800 crore in FY23 against just ₹226 crore of revenue. Reliance ultimately wrote off its entire $200 million stake, the workforce collapsed from over 800 to around 50, and the app shut down in early 2025. The lesson: pivoting toward a hotter category without a path to viable economics accelerated the end rather than saving it.