Since January I have been Head of Strategy & Growth on the founding team at ZenCabs, an all-electric cab fleet and ride-hailing app in Jammu founded by Aayush Chopra. Jammu was never a priority city for the national aggregators, so riders booked cabs over the phone and argued about fares, and drivers had no dependable flow of work. ZenCabs was built to fill that gap.
This piece walks through how the model works in numbers. Everything about ZenCabs comes from our own reporting. Everything about the wider sector is linked to its source at the end.
The operating numbers
| Metric | Where it stands |
|---|---|
| Users after six months | 30,000+ |
| Monthly active riders | 8,100, which is 27% of users |
| Rides per month | 10,000+ |
| Average order value | ₹331 |
| Monthly GMV | ₹33L+ |
| Fleet | 0 to 40 cars in three months, 10+ added a month |
| Revenue per car | ₹83K a month |
| Fleet utilisation | From 40% to 75% |
| Driver-side cancellations | From 40% to 6% |
| User cancellations | From 200 to 80 a day |
| Marketing spend | About 10% of revenue |
| Annualised run-rate | ₹4 Cr |
The figures reconcile, which is the first thing I would check in anyone else's deck. 10,000 rides at ₹331 comes to about ₹33L a month, the monthly GMV. Forty cars at ₹83K each comes to the same ₹33L, so revenue per car here means the gross value of the rides that car carried. ₹33L a month times twelve is the ₹4 Cr run-rate.
Per car, 10,000 rides across 40 cars is 250 rides a month. That is about eight rides a day, worth roughly ₹2,700 in fares.
Utilisation decides everything else
An owned fleet has costs that barely change with how much the cars move. Financing, insurance, parking and upkeep arrive every month. Revenue only arrives when a car is carrying someone. The distance between those two lines depends on the share of the day each car spends earning.
When utilisation rose from 40% to 75%, the same 40 cars could earn close to twice as much, because 75 is nearly 1.9 times 40. I could not have bought that with a marketing budget. With supply already stretched, more riders at 40% utilisation would only have produced more requests with no car to send.
So utilisation became the top number, with everything else underneath it: driver hours on the road, requests by hour and area, completed rides and the empty time between trips. We tracked 15+ KPIs in a weekly review with the founder, and every discussion had to connect back to utilisation. I have written up how that weekly review ran separately.
Fewer zones, more revenue
The largest single move was on the map. Ride data showed cars being sent to far-off pickup areas where a trip earned little and the drive out cost a lot of empty time, which also drove cancellations. I found the top 6 demand zones and phased out the far-off ones over two months. Utilisation went from 40% to 75%, revenue per car reached ₹83K a month, and user cancellations fell 60%, from 200 to 80 a day. Serving every request had felt like growth. It was lowering yield.
The 5 pm problem
We were losing evening rides with enough cars on the fleet. The hourly data showed why: driver shifts were ending between 5 and 8 pm, just as demand peaked. Moving 5 driver shifts two hours later added 200+ rides a month at no extra cost. Enough supply at the wrong hours behaves exactly like too little supply.
Two cancellation numbers, two different fixes
ZenCabs tracks cancellations from each side, and the two moved for different reasons. Driver-side cancellations were at 40%: four in every ten rides a driver accepted, the driver then cancelled. In a city where the fallback is a phone call to a driver you already know, a rider let down twice goes back to making that call.
We brought driver-side cancellations down to 6%, which is 94% ride completion, with 10+ standard operating procedures across driver onboarding and dispatch, checked every week in the review. User cancellations, the rides riders gave up on, fell from 200 to 80 a day for a different reason: the zone refocus meant fewer long waits for a car coming from far away. Retention followed both. 27% of users became monthly active riders, 8,100 people who now use the service as a habit.
Supply comes first, and it is slow to build
In a city with no good alternative, riders are cheap to find and drivers are hard to find. We designed recruitment, training and incentives for driver partners from scratch and onboarded 43 of them, while the fleet grew from 0 to 40 cars in three months, 10+ a month, backed by a market sizing and expansion case for the city. Growth stayed lean: we reached 30,000+ users with marketing at about 10% of revenue.
Keeping drivers mattered as much as finding them. Lower-performing drivers were earning too little, which hurt motivation and retention, so I redesigned the incentive cost structure. Revenue from the bottom 55% of drivers rose 70%.
Every car that arrives before there are drivers to run it pulls utilisation down. Any expansion plan in this model has to be paced by the supply of trained drivers, whatever demand looks like.
What the rest of Indian ride-hailing changed
Two things happened in the sector between 2023 and 2025 that anyone building or backing a fleet now has to price in.
Commission is turning into a subscription
Rapido moved its cab drivers to a zero-commission subscription in December 2023 and extended it to autos in February 2024, charging a flat daily fee in place of a cut of each fare. Uber followed for autos in February 2025 and said it did so to avoid being at a competitive disadvantage. Ola launched a pan-India version in June 2025, with passes at ₹67 a day covering autos, bikes and cabs.
An aggregator now lets a driver keep close to the whole fare, and an owned fleet recruits from the same pool of drivers. What a fleet can offer in return is a car the driver does not have to buy and a steady flow of bookings, and the second only holds while utilisation stays high. The subscription shift raised the bar that an owned fleet's driver economics have to clear.
BluSmart showed that a fleet is also a financing structure
BluSmart was India's first all-electric ride-hailing platform and ran its own fleet in Delhi-NCR, Mumbai and Bengaluru. It stopped taking bookings in April 2025 after SEBI's interim order of 15 April alleged that its founders had diverted ₹262 crore of loans raised to buy 6,400 electric vehicles. Most of BluSmart's cars were owned by Gensol Engineering, a listed company run by the same founders, and BluSmart later filed for insolvency.
SEBI's allegations were about the use of borrowed money, and they took down a working transport service with them. The broader point for the sector is that an owned electric fleet is a financing structure as much as a mobility business. The cars are bought with debt, utilisation services the debt, and whoever owns the vehicles controls the company. Diligence on a fleet business has to cover the lenders and the legal owner of every car, alongside the ride metrics.
How a tier-2 city behaves differently
Jammu is small next to Delhi or Bengaluru, and size cuts both ways. There are fewer competitors and a real gap to fill, which is why reliability wins over price here: a cab that turns up is worth more to a rider than a discount on one that might not. A smaller city also has fewer rides per square kilometre, so cars spend longer driving empty between trips, and utilisation has to be won through dispatch and demand planning.
The region shapes demand too. ZenCabs offers airport transfers and trips to Katra, the base town for the Vaishno Devi pilgrimage, alongside city rides. A long trip like that raises the value of a single booking and takes a car out of the city for hours, which the dispatch plan has to allow for.
What I would ask any fleet operator
- Utilisation by hour of the day, and how it has moved month on month.
- Revenue per car against the monthly cost of financing that car.
- Cancellations split by cause: no car nearby, driver declined, rider cancelled.
- The share of rides that come from repeat riders.
- How many drivers are still active 90 days after onboarding.
- Who legally owns the vehicles, and on what terms.
The same lens works on marketplaces generally. I wrote a longer list of the metrics I trust and the ones that flatter for anyone evaluating one.
