What is actually left from an order once the platform has taken its share. Worked in rupees, every assumption stated, so you can substitute your own rates. Updated July 2026.
Seven chapters, one worked model, every assumption on the page. This is not a startup guide: for that, see our complete guide to starting a restaurant in India.
A restaurant lets dine-in subsidise its delivery orders. A cloud kitchen has nothing to subsidise with, so every rupee passes through a commissioned rail. That one difference is why a restaurant template flatters a cloud kitchen.
A ₹350 order costed all the way down: commission, the GST charged on that commission, packaging and food. Every rate named, so you can put your own in and rerun it. Including the charge most models get wrong.
Commission scales with the order but your packaging does not, so ₹14 of packaging is 4% of a ₹350 order and 14% of a ₹100 one. Plus the fee your customer pays, which makes cheap orders convert worse.
The same model over a month, with fixed costs in, down to net profit. Then break-even in orders per day, and what one extra order a day is actually worth to you in rupees.
The identical order, side by side, through a platform and direct. About 1.7 times the contribution, with the costs a direct channel adds put back in honestly rather than ignored.
Chapters 6 and 7. Own kitchen versus shared space versus franchising a brand in, then the GST and FSSAI facts that change your numbers, and a monthly review that takes under an hour.
POSS pulls Zomato and Swiggy orders into one screen, holds recipe costs against each item, and reports what you actually kept per order and per brand, so this model runs on your own numbers every month.
Explore Petpooja POSSIt is the most common story in Indian cloud kitchens. Orders are up month on month, the listing looks healthy, and there is somehow less money than last quarter.
It is almost never one big leak. It is four or five small ones that each look ignorable and together exceed the margin.
Start with what a restaurant gets to forget. A restaurant that delivers has dine-in revenue paying no commission, quietly subsidising the delivery orders that do.
A cloud kitchen has no dining room. Every rupee passes through a rail that takes a cut, so the thin margin is not a side channel, it is the whole business.
Then there is the tax point most operators have backwards. Since 1 January 2022 the platform pays the 5 per cent GST on food sold through it, not you.
That is Section 9(5) of the CGST Act read with Notification 17/2021, and CBIC set it out in Circular 167/23/2021-GST. What you cannot escape is the 18 per cent charged on the commission, which a 5 per cent scheme kitchen cannot reclaim.
The third leak is the one nobody models: two charges are flat. Your packaging costs the same whatever the order is worth, and so does the fee the platform adds to your customer's bill.
₹14 of packaging is 4 per cent of a ₹350 sale and 14 per cent of a ₹100 one, which makes your average order value the number that decides everything. That is why pricing low to win volume is the trap it is.
The platform fee your customer pays works the same way. At about ₹17.58 it adds 5 per cent to a ₹350 bill and nearly 18 per cent to a ₹100 one, so your cheap item reaches checkout looking far less cheap. You never see that money, and it still costs you the sale.
Inside the guide, a ₹100 order leaves ₹25.68, so a ₹30 funded discount finishes at minus ₹4.32. You cooked, packed and delivered it, and ended poorer than if it had never arrived.
And the fourth is stopping at contribution without ever putting the fixed costs in. ₹125 an order sounds comfortable until rent and salaries arrive.
The guide runs the whole model: one order, then a month at fifty orders a day, then break-even in orders per day, then the same order sold direct instead.
If you want the operational side rather than the arithmetic, our cloud kitchen operations SOP and our note on why most cloud kitchens lose money cover it.
What the guide will not do is tell you what a good margin looks like. The figures circulating online for Indian cloud kitchen margins and failure rates trace back to vendor blogs, not studies.
It shows you how to compute your own, and states every assumption so you can disagree with it.
Identical commission, identical food cost percentage, identical packaging. The only thing changing is the order value. Assumptions: 24% commission, 18% GST on it, ₹14 packaging, 32% food cost. Swap in your own from your commission statement.
What is left from a ₹350 order after 24% commission, the 18% GST on it, ₹14 packaging and 32% food cost. That is contribution, not profit: rent and salaries still have to come out of it.
Source: Worked model in this guide. Every rate stated and changeableWhat ₹14 of flat packaging costs you on a ₹100 order. On a ₹350 order the same packaging is 4%. Nothing you do to food cost fixes this, because packaging does not scale with the order.
Source: Worked model in this guide, at ₹14 of packaging per orderSince 1 January 2022 the aggregator, not you, pays the 5% GST on food sold through its platform. Your platform revenue carries no output GST for you. The 18% on commission is still yours, and is not reclaimable.
Source: Section 9(5), CGST Act 2017 with Notification 17/2021-Central Tax (Rate), clarified in CBIC Circular 167/23/2021-GSTThe instinct is understandable and the arithmetic is against it. Flat ₹14 packaging is 4% of a ₹350 sale and 14% of a ₹100 one, and the platform adds a flat fee to your customer's bill on top. The cheap order you win is the one you earn least on.
Since January 2022 the platform pays that, not you. An operator still taking it off their own revenue is understating their profit and may be pricing higher than they need to.
It is charged on the commission and a 5 per cent scheme kitchen cannot reclaim it. On a ₹350 order that is ₹15.12 of real cost, and it is missing from most home-made models.
A funded discount comes off contribution, not off commission. Run the order through the model with the discount in first. Our note on staying profitable on Swiggy and Zomato commission covers the negotiation side.
₹125 an order feels comfortable until rent, salaries and platform advertising arrive. Contribution pays for those. Whatever survives them is profit, and the break-even point is where the two meet.
A direct order carries about 1.7 times the contribution. Averaging them hides which channel is paying you and makes any decision about either one guesswork.
A fitted-out kitchen sets a high, rigid break-even before you have sold anything. A shared kitchen costs more per order and can be walked away from. Most first kitchens should start there, and our guide to starting a profitable cloud kitchen covers the setup choices.
A direct order is worth about ₹92 more, roughly 1.7 times the contribution. The guide puts the costs a direct channel adds back in, including the GST that becomes yours again.
| Line | Through an aggregator | Direct |
|---|---|---|
| Order value | ₹350.00 | ₹350.00 |
| Platform commission at 24% | ₹84.00 | nil |
| GST on that commission | ₹15.12 | nil |
| Payment gateway at 2% | included above | ₹7.00 |
| Packaging | ₹14.00 | ₹14.00 |
| Food cost at 32% | ₹112.00 | ₹112.00 |
| Contribution per order | ₹124.88 | ₹217.00 |
| Who pays the 5% GST on food | The platform, under s.9(5) | You do |
| Who handles delivery | The platform | You, or someone you pay |
Swipe the table sideways to see the direct column.
Download the free guide and run the model on your own commission rate this week.
Petpooja is India's leading SME business software suite, trusted by 1,50,000+ businesses across restaurants, retail, healthcare, manufacturing, and more. From billing and payroll to task management and procurement Petpooja helps Indian businesses run better, every day.
Petpooja POSS pulls Zomato and Swiggy orders into one screen, holds recipe costs against each item, and separates direct from aggregator sales, which is exactly what this model needs each month.
