Your Bestselling Dish Is Probably Losing You Money. Here's the ₹20,000 Business Nobody in India Is Running Yet.
There's a chicken biryani in a 40-seat restaurant somewhere in Indore that sells 60 plates a day on Swiggy. The owner is proud of it. It loses him ₹11 a plate.
He doesn't know this because nobody has put the four numbers in the same room: what the dish costs to make, what the platform deducts, what he funds in discounts and what the packaging adds.
The food cost is in his head. The commission is in a settlement PDF. The discount is a checkbox he ticked eight months ago. The packaging is a monthly bill from a supplier.
Put them together and the biryani is a charity operation running 1,800 times a month. This is the gap. And it's a business.
The setup: everyone has the data, nobody reads it
Every restaurant listed on Swiggy or Zomato is sitting on a pile of exportable data:
The owner opens maybe two of these. He looks at total sales and the amount that hit his bank. If the second number is smaller, he assumes it's “commission” and goes back to the kitchen.
The margin for error is nearly zero.
Base commissions are widely reported in the 17–28% range. Add GST on commission, payment collection fees, packaging deductions, restaurant-funded discounts and ad spend: effective deductions commonly land in the 25–35% band.
What the audit actually is
Call it a Profit Diagnostic. It is two to three days of work: take the exports, run the analysis and hand back a report that answers six questions the owner has never been able to answer.
Which popular dishes are actually unprofitable?
Calculate true contribution per dish: menu price minus food cost, packaging, effective platform deduction and that item's share of discount funding. Plot volume against contribution and classify every item as a Star, Plowhorse, Puzzle or Dog.
Almost every menu has two or three high-volume, negative-or-near-zero-margin items.
What should be bundled?
Use basket analysis to find items that appear together more often than chance would predict. If 34% of biryani orders include raita and a soft drink, a ₹319 combo can raise AOV, reduce platform fees as a percentage and move a high-margin beverage.
Beverages and desserts are where the margin hides. Most menus bury them.
Where is discounting bleeding out?
Quantify discount funding by item and platform, then compare it with incremental orders. Owners switch on “Flat ₹125 off” during a slow month and forget to switch it off — even on items that were already thin.
Frequently, a discount bought orders from customers who would have ordered anyway.
What are customers actually complaining about?
Structure the review corpus with sentiment and theme extraction. “Food was bad” becomes a useful operating signal: gravy leaks after 9 PM, portions look smaller than the photo, or paneer gets 3.1 stars while chicken gets 4.4.
These are fixable problems with named owners and deadlines.
Which descriptions and photos cost conversions?
On a delivery app, the photo is the product. Rewrite the top 20 item descriptions and create a shot list for a two-hour photo session. Bright, top-down images and appetising copy beat inventory labels and stock photography.
“Slow-simmered tomato-butter gravy” sells better than “Paneer Butter Masala — 300g.”
Where does AOV grow?
Test add-on prompts, price ladders, minimum-cart thresholds, upsizing and a ₹49 dessert positioned against a ₹199 one. A ₹40 AOV increase on 1,500 monthly orders is ₹60,000 in incremental revenue.
Most of that flows to contribution because fixed costs are already paid.
The math that makes this an easy sell
Here's an illustrative case: a mid-size outlet doing roughly ₹6 lakh a month across both platforms.
| Finding | Monthly impact |
|---|---|
| Reprice 3 loss-making items, delist 4 dead ones | ₹22,000 |
| Kill two stale discount campaigns | ₹31,000 |
| Introduce 3 combos, lift AOV by ₹35 | ₹42,000 |
| Fix packaging on the top leaking item, cut refunds | ₹8,000 |
| Rewrite + rephotograph top 15 items | ₹18,000 |
| Total illustrative impact | ₹1,21,000 |
Illustrative, not a promise. Real numbers depend entirely on the outlet — some will be half this, a few will be double.
Now ask for ₹20,000.
It isn't a cost. It's a fee against a number the owner can see on their own dashboard within six weeks. You are not selling analytics. You are selling found money.
Why this business exists now
The data got digital
Everything is exportable now. There is no manual data-collection phase.
The analysis got cheap
AI can clean exports, categorise reviews, run basket analysis and draft the report. You supervise and sanity-check.
The pain got sharp
With effective deductions in the 25–35% band, restaurants can no longer survive on instinct alone.
Work that would have cost ₹3 lakh as a Big Four engagement can now be delivered for ₹20,000 at a healthy margin. That's a new market opening.
What this looks like as a business, honestly
| Offer | Price / arithmetic |
|---|---|
| Initial profit audit | ₹15,000–₹25,000 |
| Ongoing monthly monitoring | ₹5,000–₹10,000 |
| 3 audits each month | 3 × ₹20,000 = ₹60,000 |
| 15 retainer clients | 15 × ₹7,000 = ₹1,05,000 |
Roughly ₹1.6 lakh/month
The audit is a one-time event; monitoring is the annuity. Monthly send a short dashboard: contribution by item, discount spend, AOV trend, rating movement, new complaint themes and two or three actions.
Fifteen retainers is not a fantasy. That's fifteen restaurants in one city — a single decent food cluster in Pune, Jaipur or Kochi has more than that.
Startup cost: effectively your laptop.
Add an AI subscription and a spreadsheet template. No inventory, office, team or licence. Your first client's fee covers everything you spent.
Your real moat is the format: a report so clear that a 52-year-old owner can reach page 2 and say, “Arre, the biryani is losing money?”
How to get your first three clients this month
Pick one cluster
One neighbourhood, one cuisine, 20 outlets within 3 km. Depth beats spread — your third client in a cluster comes from the second one talking.
Do one audit free
Pick a restaurant you already eat at. Ask for exports, deliver the full report and ask for nothing. You now have a case study with real numbers and an owner who can vouch for you.
Lead with a finding
Don’t say “I do AI-powered restaurant analytics.” Say: “I pulled your public reviews — 38 of your last 100 mention delivery packaging, concentrated on gravy items after 9 PM. Want to see what else is in there?”
Charge low, then hold
Charge ₹15,000 for clients one through three, ₹20,000 after two case studies and ₹25,000 once you have a testimonial with a rupee figure in it.
Sell the retainer at the debrief
Offer ongoing monitoring while the owner is looking at the ₹1.2 lakh opportunity. The audit is the one-time event; the dashboard is the annuity.
The objections you'll hear
“I don’t know restaurants.”
You don’t need to. You need to know unit economics, and the owner supplies the domain knowledge. Your value is being the only person in the room who looked at all the data at once.
“Won’t Swiggy and Zomato just build this?”
Platforms have dashboards, but no incentive to tell a restaurant that ad spend is wasted or discount funding bought no incremental orders. You are on the restaurant’s side of the table.
“What if the owner does nothing with the report?”
Some will. That is why the retainer exists and why the report ends with three actions, not thirty insights. Rank findings by rupees, cap the list at five and put a name and date next to each.
“Isn’t the market small?”
India has lakhs of restaurants listed on delivery platforms. You need fifteen retainers — fifteen restaurants in one city. A single food cluster in Pune, Jaipur or Kochi has more than that.
Start with one
The barrier isn't capital or credentials. This work sits in the gap between “restaurant person” and “data person,” and very few people stand in that gap.
Pick one restaurant you like. Ask for the last three months of exports. Spend a weekend building the report you'd want if it were your money on the line.
Then watch the owner's face when they get to the biryani. That expression — that's the whole business.