How to start a cloud kitchen with a team

What you need before you open

Every team that opens one of these ends up with the same list. What it adds up to is not a fact about the business — it depends on your city, on what you can find second-hand, and on how much of it somebody in the team already owns. So the useful question is not the total. It is which of these lines you can already cover between you, and that is a question you can answer in an evening.

• Kitchen deposit and first month rent

  A 150–250 sq ft shared cloud-kitchen bay near an office corridor, charged monthly, usually with two months taken as deposit up front.

• Commercial equipment

  Four-burner range, exhaust, deep fryer, refrigeration, work tables. Second-hand equipment from a closed kitchen cuts this by a third.

• FSSAI State licence and GST registration

  State licence, not basic registration — aggregators require it, and turnover here crosses the basic tier's ceiling almost immediately.

• Menu photography and brand pages

  The single highest-return spend in this business. Bad photos cost you more than bad food, because nobody orders the food.

• Opening stock and packaging

And every month after that

• Kitchen rent

• Provisions

  Ingredients only, costed per order at a run rate of 900 orders a month.

• Packaging

  Costed per order. Leak-proof containers are emphatically not the place to save money.

• Gas, electricity, water

• Aggregator ads

  Optional in month one, unavoidable by month three if you want to hold a top-ten position in your locality.

Licences and registrations

• FSSAI State Licence

  Owned by: Supply & Money

  Every country licenses businesses that cook for sale; in India that is FSSAI. The State Licence is required once turnover passes the basic tier's ceiling, which a cloud kitchen does in its first quarter. Apply on foscos.fssai.gov.in with the kitchen lease, a layout plan, a water test report and the partners' ID proofs; the fee is modest and depends on the term you choose. Swiggy and Zomato will not activate a listing without the number, so start this before you sign the lease, not after.

• GST registration

  Owned by: Supply & Money

  Delivery platforms everywhere settle to a tax-registered business; in India that is GST. Mandatory in practice: the aggregators deduct TCS and require a GSTIN to settle payouts. Restaurant service is taxed at 5% without input credit. Register on gst.gov.in before your first payout cycle, or your money sits held.

• GHMC Trade Licence

  Owned by: Supply & Money

  Every municipality licenses food businesses trading from a premises; in Hyderabad that is the GHMC trade licence. A trade licence for the kitchen premises, applied for through the GHMC citizen portal for your circle. Most shared cloud-kitchen operators hold a building-level licence already — ask the landlord for a copy before you apply separately, since it can cover you.

• Fire safety NOC

  Owned by: Head Cook

  Required for commercial cooking premises. In a shared kitchen the operator normally holds it building-wide; get written confirmation, because the absence of one is what closes kitchens during inspection drives.

Who does what, and what share they hold

Shares are of revenue after costs, agreed in writing before the work starts.

• Head Cook

  Owns the menu, the prep list, and whether order 200 tastes like order 2.

  54 hrs/week

  34%

• Kitchen & Dispatch

  Second pair of hands on the line, packs every order, and keeps the rider handover under two minutes.

  45 hrs/week

  24%

• Listings & Growth

  Runs the Swiggy and Zomato pages, the photography, the pricing and the ad spend.

  25 hrs/week

  24%

• Supply & Money

  Buys provisions, tracks cost per plate, and reconciles the aggregator payouts.

  20 hrs/week

  18%

The six-week plan

1. Week 1

   1. Pick two brands, not one

      A single kitchen should list as two or three separate brands on the aggregators — biryani under one name, rolls and wraps under another. It costs nothing extra, the same prep line serves both, and it doubles the surface area you occupy in search results. What it must not do is double your ingredient list: design the second brand from things the first one already stocks.

   2. Choose a locality before you choose a kitchen

      Aggregator reach is radial and brutal: past about 6 km your food arrives cold and your rating falls. Pick the delivery zone first — an office corridor for weekday lunch volume, a late-closing nightlife strip for the late-night rush — then find a kitchen inside it. Open the apps at 1 pm and at 9 pm in your target area and count how many kitchens already sell your dish. Twelve competitors is a functioning market; two is a warning.

   3. Cost the plate before you love the recipe

      Work out the landed cost of every single dish: ingredients, packaging, gas, and then the aggregator commission of eighteen to twenty-five per cent that comes off the top of the menu price before you see any of it. Once all four are subtracted, roughly a fifth of what the customer paid is left, and THAT is your actual margin — not the number on the menu. Set a floor in rupees per dish and hold it without sentiment. If a dish cannot clear the floor, it does not go on the menu, however much you like cooking it and however good it is.

2. Week 2

   1. Sign the kitchen and register the business

      Shared cloud-kitchen bays near an office corridor run 150 to 250 square feet, charged monthly with typically two months taken as deposit. Before signing anything, confirm IN WRITING that the building's own FSSAI, fire NOC and trade licence actually cover your bay — several operators let bays that their licences do not extend to, and you find out when an inspector does. Start your own FSSAI State Licence and GST registration in the same week: both take two to three weeks, and both block your listings from going live.

   2. Build a menu of fourteen dishes, maximum

      Every additional dish adds prep, stock and waste. Fourteen items that share a base — the same gravy, the same marinade, the same rice — will outsell thirty that do not, because a short menu is a fast kitchen. Aim for a 28-minute average prep time; the aggregators rank on it, and customers cancel over it.

3. Week 3

   1. Photograph the food properly

      Budget one full day and a real photographer. Every dish, natural light, real portions in the actual packaging you will send them out in — not a styled plate the customer will never receive. This is the highest-return line in the whole build and it is not close: on an aggregator nobody tastes anything before ordering, so the photograph IS the product. Kitchens with phone snaps taken under a tube light lose to identical food shot well, every single time, and they never find out that is why.

   2. Go live on both aggregators

      Onboarding needs the FSSAI number, the GSTIN, a cancelled cheque, the menu with prices and the photographs. Expect 5–10 days from submission to live. List on Swiggy and Zomato both — the overlap in customers is smaller than you would think, and depending on one platform hands it your pricing.

4. Week 4

   1. Run the first hundred orders at full attention

      Both partners on the line, no shortcuts, every order photographed before it is sealed. Your rating in the first fortnight sets your ranking for months, and a 4.6 that starts low is far harder to lift than one that starts high. Refund any complaint immediately and without argument for the first two weeks; it is the cheapest reputation you will ever buy.

   2. Instrument the two numbers that matter

      Average preparation time and order accuracy. Both are visible in the partner dashboards and both drive your ranking directly. If prep time drifts past 30 minutes, the menu is too long or the line is understaffed — fix that before you spend a rupee on ads.

5. Week 5

   1. Start advertising, but only on winners

      Once you have thirty ratings and not before, put a fixed monthly budget behind the three dishes with the best combination of rating and margin. Never advertise a dish you are unsure about: paid traffic to a weak dish buys you bad reviews at a premium, and those reviews outlive the campaign by months. Track cost per acquired order rather than impressions, set a ceiling for it against your per-dish margin, and switch the campaign off the moment it goes above — a campaign that acquires orders for more than they earn is a machine for losing money quickly.

   2. Watch the payout cycle, not the order count

      Aggregators settle weekly, net of commission, TCS and any promotional cost, usually 7–10 days in arrears. A kitchen can look busy and still run out of cash because provisions are paid daily and revenue arrives weekly. Keep one month of provisions as float and reconcile every payout against the order list — deductions are common and are not always right.

6. Week 6

   1. Add the third brand, or the second shift

      By now you know which of your two brands earns. The cheapest growth is a third brand on the same prep line, or extending into late-night when the kitchen is already staffed. Both add revenue without adding rent, which is the only reason this business model works at all.

   2. Close the month and pay the crew

      Log every rupee in and out, reconcile both aggregator statements, then split on the percentages in the signed agreement. The Head Cook carries 34% because they carry the hours and the consistency risk. Pay on a fixed date, including the months the number disappoints — that is when the discipline actually matters.

The name comes with the network

Teams that start this blueprint open as Junto Kitchen, numbered in the order they open — the first as Junto Kitchen #1, the next as Junto Kitchen #2, and yours takes whichever number is next the day you start. One name, many small teams, each with its own Pact and its own earnings. Packaging, provisions and aggregator onboarding through the network’s contacts.

• A name that compounds

  You open as Junto, numbered in the order teams open. Customers who trust one team trust the next, and every team that opens adds to what the name is worth.

• Suppliers who take your call

  Procurement contacts the network has already negotiated with. You buy at network prices from day one, not at walk-in prices.

• Shared services at network rates

  A call centre that answers your customers, and the back-office pieces no three-person team should have to build alone — run once for the whole network.

Prefer your own name? Opt out when you start. Everything else — the blueprint, the Pact, the ledger — works exactly the same.

Questions people ask

How much does it cost to start a cloud kitchen?

Five lines, in the order they matter: the deposit and first month on a shared kitchen bay, commercial equipment, one proper menu photography day, opening stock and packaging, and the FSSAI and GST registrations. The bay deposit and the equipment together are most of it, and everything else is small beside them. Buying equipment second-hand from a kitchen that has closed is the single biggest saving available here and it is usually about a third off. This is the most expensive thing in the launch set to start, which is worth knowing before you pick it.

Do I need an FSSAI licence for a cloud kitchen?

Yes, and specifically a State Licence rather than basic registration, because a cloud kitchen crosses the basic tier's turnover ceiling within its first quarter. Swiggy and Zomato will not activate your listing without the number, so this is not paperwork you can defer. Apply through foscos.fssai.gov.in with your kitchen lease, layout plan, water test report and ID proofs; it takes about two to three weeks.

How much commission do Swiggy and Zomato take?

Typically eighteen to twenty-five per cent of order value, plus payment gateway charges and any promotional spend you opt into. Settlement is weekly and roughly seven to ten days in arrears, which matters more than it sounds — your ingredients are bought daily in cash. Price every dish against the post-commission number rather than the menu price: once ingredients, packaging and commission come off, roughly a fifth of what the customer paid is actually yours.

How many orders a month does a cloud kitchen need?

Around 380 to 420 orders a month covers rent, provisions, packaging and utilities. A four-person crew starts earning properly at roughly 900 orders a month. The usual way to reach that volume is not more marketing but more listings: running two or three separate brands off one prep line adds orders without adding a single rupee of rent, and it is what most kitchens that survive their first year end up doing.

Is a cloud kitchen better than a restaurant?

It is cheaper to start and far cheaper to be wrong in. There is no dining room, no front-of-house staff and no high-street rent, so the build is a fraction of what a restaurant costs and closing one loses you a bay deposit rather than a fit-out. The trade-off is real and permanent: you own no customer relationship at all. The aggregator does, it takes roughly a fifth of every order, and it can change your ranking whenever it likes without telling you.