There is no shortage of restaurant software aimed at the Indian market. Petpooja runs at over 55,000 outlets. Posist and Restroworks dominate the chain segment. Marketman, Wisk, and Supy land from overseas. Layer in Zomato and Swiggy aggregator dashboards, and a small operator can reasonably feel they have already bought too much software, not too little.
So why are most independent cafés and small restaurant brands in India still running their actual operations out of a WhatsApp group, a paper indent book, and a heavily massaged Google Sheet?
Because back-of-house is not one job, and packaged products are good at one specific slice of it. This is a field guide for picking software if you run one to five F&B outlets in India, lean specialty, and care about margin discipline. We will cover what the packaged products actually do, the four layers of an F&B stack, where the packaged products end, and what to build instead of replace.
What packaged products actually solve
Walk a Petpooja, Posist, or Restroworks demo and you will see roughly the same surface. The category has matured and converged.
- Billing and KOT. GST-compliant invoices, kitchen order tickets, dine-in / takeaway / delivery splits, table management.
- Aggregator integrations. Direct API into Zomato, Swiggy, and ONDC partners. Menu sync. Order auto-acceptance.
- Payment reconciliation. UPI, cards, cash. End-of-day shift settlement.
- Item-level inventory. Stock in, stock out, reorder thresholds, basic recipe cards (this dish uses 250g of that raw).
- Aggregator commission tracking. What Zomato actually pays out vs the order value.
- Reports. Daily sales, top items, hourly heatmap, payment mode split.
- Loyalty and CRM. Customer phone-number-based.
If you run one outlet, this is genuinely enough. Pay the per-outlet subscription, train the captain and the cashier, and run the business. The pain you feel will be operational discipline, not software.
The wheels start coming off when you cross three outlets, run any kind of central kitchen or commissary, or run a menu where ingredient yield actually matters — meaning specialty coffee, baking, butchery, anything with prep beyond plating. That is where the four-layer reality kicks in.
The four layers of an F&B back of house
The mental model that has held up across every F&B engagement we have looked at:
Layer 1 — Counter and KOT. The cashier billing, kitchen tickets, and aggregator reconciliation. This is what packaged POS products are best at, and the gap between them and a custom build is enormous in the wrong direction. Do not rebuild this layer.
Layer 2 — Outlet inventory. What is on the shelf, what reorder is due, what spoiled. Packaged products have a module for this. It works, with discipline. The discipline is the hard part.
Layer 3 — Central kitchen and procurement. What the commissary produced today, what indent each outlet raised, what supplier delivered, what the supplier was supposed to deliver, and what the variance was. The packaged products either do not have this layer or have a notional version that no one uses past the demo. This is where most chains end up running a Google Sheet.
Layer 4 — Owner cockpit. Daily P&L by outlet. Wastage by station and by shift. Cost-of-goods drift across the menu. Comparable-outlet performance. The board view. The packaged products give you reports; they do not give you a cockpit. There is a real difference.
The mistake we see most often is treating layers 3 and 4 as features the POS will eventually grow into. They will not. Those are different products with different rhythms and different users, and the right play is usually to keep the POS for layers 1 and 2 and build thin, focused surfaces for the other two.
Where the packaged POS runs out
Five things start to break the moment you push a packaged restaurant POS past one or two outlets:
1. Recipe cards do not handle yield variance. A packaged recipe module assumes 1kg of raw butter chicken gravy is always 1kg, regardless of which kitchen made it. In real kitchens, ghee absorption shifts, water reduction shifts, and the same recipe yields differently across cooks. For specialty coffee this is brutal — bean yield drifts with grind, dose, roast date, and humidity. The recipe card has to be a calibrated, dated artifact, not a one-time entry.
2. Central kitchen indent is invisible. You have a commissary that produces base sauces, prepped vegetables, and mise-en-place for three outlets. Each outlet raises an indent every evening. The commissary fulfils next morning. The packaged POS does not see any of this. Most operators run it on WhatsApp and a paper indent book, and the variance between indented vs received vs billed quietly eats two to four points of margin a month.
3. Wastage is captured nominally or not at all. Packaged products have a "wastage" field in inventory adjustments. Nobody fills it in. The reason is workflow — wastage happens during prep at 5pm, the inventory module is on a desktop in the manager's cabin, and the prep cook is not going to walk over. The capture surface has to live where the wastage happens, on a phone or a station tablet, in 10 seconds, with a photo.
4. Multi-outlet reconciliation is a spreadsheet. Three outlets, three Petpooja accounts, one P&L. Most packaged products do roll up reports across outlets, but the moment you have one outlet on a different brand, a cloud kitchen on a different POS, or a catering vertical that bypasses the till, the consolidated view falls apart. We have seen four-outlet brands running an end-of-day spreadsheet that takes the manager two hours every night.
5. Supplier and FSSAI traceability is a folder. Most packaged products give you a "documents" tab and a place to upload supplier invoices. The real workflow — supplier batch number captured at receipt, linked to incoming SKU, tied to which dish it ended up in, available for an FSSAI inspector in 30 seconds — does not exist. It is buildable in a week. It almost never gets built.
What to build instead of replace
The pattern that has worked for the F&B operators we have shipped to is simple: keep the packaged POS, build a thin custom layer on top of it that handles the four problems the POS does not.
A typical first version, scoped to ship in 6 to 10 weeks:
- Central kitchen indent flow. Outlet manager raises the indent on a phone screen at evening close. Commissary fulfils next morning, marks delivered quantity, captures variance if any. Auto-posts the deduction to the receiving outlet's inventory in the POS via API.
- Prep wastage capture. A station tablet or shared phone with a 3-tap flow — what got wasted, why, who logged it. Photo optional. End-of-day report by station and shift, surfaced to the head chef on WhatsApp.
- Daily outlet P&L. Net sales pulled from POS. COGS pulled from inventory consumption. Rent, payroll, and aggregator commission overlaid from a small static config. One number per outlet per day, in a 9am Slack or WhatsApp message to the owner.
- Recipe yield calibration. A weekly job where the head chef updates yield factors per dish per kitchen — coffee dose, bread yield, sauce yield. The cockpit highlights menu items where actual cost has drifted more than 5% from theoretical.
- Owner cockpit. A single screen on a phone with daily P&L, wastage by station, top three drift dishes, indent variance. Updated overnight. Read-only. Loved by founders.
That is roughly INR 6–15 lakh of work depending on outlet count and integration depth. It sits on top of whichever POS the operator has already paid for, talks to it via API where one is exposed, and gives the operator what the packaged product was never going to give them.
We covered a version of this build for a Pune coffee house at our Soch et Latte case study — replacing a tablet POS, a paper indent book, and three WhatsApp groups with one system the staff actually wanted to use, in six weeks.
A worked example
A specialty café group, 3 outlets in Pune and Mumbai, INR 3.5cr ARR, running on Petpooja with a Google Sheet for the central kitchen.
The owner spends about 6 hours a week reconciling the sheet. Two of three outlets have monthly food-cost variance of 4–6% that nobody can explain. Wastage is "logged" in a paper book that gets thrown out on Saturday.
Scope of a custom layer on top of Petpooja:
- Indent flow for the central kitchen feeding 3 outlets
- Wastage capture on a shared shop-floor phone
- Daily outlet P&L with COGS, rent, payroll overlay
- Owner cockpit with the four numbers the founder actually checks
- FSSAI document trail tied to supplier deliveries
Build cost: ~INR 8 lakh fixed-price. Timeline: 8 weeks. Retainer after launch: ~INR 35,000/month for changes and support, in line with what we describe in our internal tools service. Total year-one outlay including the retainer: ~INR 12.2 lakh.
What it pays back: 6 hours a week of the founder's time, ~2 points of margin recovered from variance and wastage discipline, and a real document trail at FSSAI audit. The Petpooja subscription stays. None of the discipline of the cashier flow gets touched. The operator stops thinking about software for the next 18 months.
When packaged is still the right answer
To be fair to the category — there are three situations where a custom layer is a mistake.
One outlet, standard menu. Petpooja is the right answer. Anything we build on top is overkill.
No central kitchen. If every outlet preps independently and you have no commissary or production kitchen in the loop, the indent layer disappears. The remaining problems — wastage and the cockpit — are real but smaller. A small process automation engagement wiring Petpooja to a Google Sheet dashboard might be all you need.
Pre-revenue or pre-traction. If the brand is still finding its menu and footfall, software discipline is not the bottleneck. Get the menu right first.
In every other case — multi-outlet, central kitchen, margin discipline a real concern — the right play is the four-layer model. Keep the POS. Build the surface it skips. Stop running your operation out of a WhatsApp group.
If you are a small F&B group somewhere in this picture and the central kitchen, wastage, or owner cockpit problem is starting to feel real, come and talk to us. We have shipped this pattern enough times to know which corners to cut and which to take seriously.