No government body or industry association in India publishes a restaurant failure rate. Not the food services associations, not the ministries, not the ratings agencies. Anyone quoting you a precise Indian figure is quoting something nobody measured.
The largest study on the question tracked 2,439 restaurants in Columbus, Ohio, between 1996 and 1999. It found 26% failed in year one, 19% in the second and 14% in the third. Cumulatively, 59% were gone inside three years.
So about one in four in year one, and three in five within three years, on American data that is now nearly thirty years old. The famous 90% figure is not research at all.
We work with restaurant kitchens across India every day. The ones that close rarely do so for the reasons owners fear when they sign the lease.
That matters if you are opening in Kharadi or Vastrapur next quarter. The number you were frightened by is wrong, and the real risks sit somewhere else.
Key Takeaways
- India publishes no restaurant failure rate. Treat every Indian percentage you read as unsourced until someone names the study.
- The strongest available research found 26% first-year failure and 59% over three years, in Columbus, Ohio, 1996 to 1999.
- A separate study of 81,000 restaurants found independent restaurants failed slightly less often in year one than other service businesses.
- The 90% claim traces to a 2003 advertisement and a television show, not to any study.
- What India does publish is market size, employment and GDP share, all of which are growing.
How Many Restaurants Actually Fail?
The honest answer has three parts, and only one of them is usable.
| Claim | Where it comes from | Can you rely on it? |
|---|---|---|
| 90% in year one | A 2003 advertisement and a TV programme | No. No study has produced it |
| 60% in year one, 80% by year five | Repeated across vendor blogs with no attribution | No. No named study, sample or method |
| 26% year one, 59% over three years | 2,439 restaurants, Columbus Ohio, 1996 to 1999 | Yes, with the date and the geography attached |
| 17% year one | 81,000 restaurants, US Bureau of Labor Statistics census | Yes, same caveat |
| Any India-wide rate | Nothing. It is not published | No, because it does not exist |
The 90% claim has a traceable origin, and it is not research. It comes from a 2003 American Express television commercial featuring chef Rocco DiSpirito, tied to his NBC show The Restaurant.
That advert is what prompted the Columbus research in the first place. H.G. Parsa, the Ohio State researcher behind it, put his finding bluntly. The 90 percent figure, he said, “seems to be a myth, a myth that is harmful to the restaurant industry.”
The second piece of work surprises people. It covered 81,000 full-service restaurants in a Bureau of Labor Statistics census of the western United States, run over two decades.
It found that 17% of independently owned restaurants failed in year one, against 19% for every other kind of service startup, which means restaurants did slightly better than the businesses nobody worries about.
Why India Has No Published Restaurant Failure Rate
This is not an oversight. It is a measurement problem.
Tracking a failure rate means following one year’s intake of new restaurants and recording which ones die. India has no such register.
Most of the market is unorganised, licences record openings rather than closures, and an outlet that quietly stops trading leaves no record at all.
What the sector does publish is size, not survival. The India Food Services Report puts the market at ₹5.69 lakh crore in FY24, heading for ₹7,76,511 crore by FY28. It employs 85 lakh people directly and accounts for 1.9% of India’s GDP.
| What India publishes | What India does not publish |
|---|---|
| Market size and forecast | First-year failure rate |
| Outlet counts by format | Three-year or five-year survival |
| Direct employment | Cohort tracking of new openings |
| Share of GDP | Average lifespan of an outlet |
Read that table the right way round. A sector growing at 8.1% a year is not one where nine in ten businesses die. It is one where nobody counts the ones that do.
One Indian closure figure does exist. In FY21 the sector shrank 53%, and over 25% shut permanently, costing close to 2.4 million jobs.
That was a pandemic, not a normal year. Quoting it as India’s failure rate would be the same borrowed-number trick this article is arguing against.
Why Do Most Restaurants Fail in Their First Years?
Since the rate is unmeasured, the causes are what actually help you. These are the eight that come up most often, and the order matters less than the fact that most closures involve several at once.
1. Insufficient Market Knowledge
Owners commit their savings before testing whether the demand is there. They misjudge what the return will be, skip the survey work, and open on instinct.
The research is not a one-time exercise either. Tax rules change, delivery behaviour shifts, a competitor opens across the road. A bar beside a school and a bakery far from any footfall both fail for the same reason, which is that nobody checked first.
The cheapest version of this check costs a fortnight. Stand outside your proposed site at 1pm on a weekday, then again at 8pm. Count the people who pass, then count how many of them walk into the eateries that are already there. That number will tell you more than any market report you buy.
2. Poor Location
Rent is the line you cannot renegotiate later. Sign for a site at, say, ₹1,15,000 a month because the frontage looked impressive, and you have fixed your break-even before serving a single cover.
Connectivity, parking, and the catchment within a ten-minute walk decide more than the interiors do. Our guide to choosing a restaurant location covers what to check before signing.
3. Menu and Prices
A menu that runs to eleven pages ties up cash in ingredients that move twice a week. Every extra dish adds stock, waste and kitchen time.
Pricing gets set by looking at the shop next door rather than by costing the recipe. Work out the food cost on each dish first, then price it, then cut whatever cannot carry itself.
A profitable menu is usually a short one. How it is laid out matters too, because the layout decides which dishes get noticed at all.
4. Poor Staff Management
Kitchen churn is the cost nobody budgets. Every departure means recruiting, training, and a period where the food is inconsistent.
Underpaying and overworking a team saves money on paper and loses it on the floor. Reasonable benefits cost less than replacing a chef twice a year.
Rosters are the other half of it. Staff a Tuesday lunch like a Saturday dinner and you burn wages on an empty floor. Do the reverse and you burn goodwill. Both mistakes are visible in your own sales data long before they show up in the bank balance.
5. Overlooked Expenses
Owners budget for rent, salaries and raw material, then get caught by everything else. Licences, insurance, gas, electricity, repairs, delivery packaging, card charges. Every one of them is operating cost, and together they decide whether a good sales month clears anything.
Fuel and power move without warning, and they land on a line most owners never modelled. Then there is the money that leaks quietly, through wastage, over-portioning and pilferage.
6. Poor Customer Experience
One bad Saturday reaches a few hundred people before Monday. A slow kitchen, a rude greeting, a cold plate, and the review is written in the car park.
Handling that badly makes it worse. Answering it in public, quickly and without argument, is worth an hour of any owner’s time.
7. Lack of Marketing
Good food does not market itself, and the belief that it does closes restaurants. If nobody within three kilometres knows you opened, the quality of your biryani is irrelevant.
A website you own beats a listing you rent, and good photographs cost less than a week of discounting.
8. No Business Analysis
This is the one that ties the other seven together. Owners who cannot say which dish made money last month are guessing, and guessing gets expensive.
“The main reason that the mortality rate of restaurants is so high today is because owners are not looking at the numbers. All the answers lie in the dashboard of your POS system.” Pravesh Pandey, Chief Instigator and Belief Officer, BYG Brewski
The reports worth watching are the boring ones: item-wise sales, wastage, and daily variance between what sold and what left the store.
None of that needs a finance background. A restaurant POS records it as a by-product of billing, which means the answer is usually already sitting in the dashboard nobody opens.
What are the Early Warning Signs?
Closures are slow. The signals show up months before the shutters do, and they are all countable.
- Weekday covers falling while weekend covers hold. Regulars leave first and they leave quietly. A weekend-only crowd is a restaurant living on discovery traffic rather than repeat custom.
- Food cost climbing without a price change. If your food cost ratio moved from 30% to 36% and nothing on the menu got dearer, the leak is portioning, wastage or stock walking out.
- Paying suppliers later each month. The date you settle vendor bills is the most honest indicator on the premises. When the 7th becomes the 15th, working capital is already thin.
- Discount-dependent delivery volume. If orders collapse the week you stop a promotion, you are buying revenue rather than earning it.
- Staff leaving without being replaced. Running short-handed to save wages is a decision most owners make once and then cannot reverse, because service slips and covers follow.
Any one of these is manageable in isolation. Two or three at once, over a quarter, is the pattern that ends in a closed shutter.
Where the Margin Actually Dies
Instead of a failure rate, use arithmetic. This is where a rupee of revenue goes in most Indian kitchens.
Those ranges are planning conventions rather than research findings, and they move by format. A cloud kitchen carries almost no rent and far more commission. A fine-dine floor in Lower Parel is the reverse.
The delivery line is the newest pressure and the one that catches operators out. Commission is charged on the menu price. So a discounted dish can leave the kitchen at a loss while the sales figure still looks healthy. We have set out how those deductions stack up in a separate commission breakdown.
Example (illustrative). Say a QSR in Koramangala sells a ₹340 rice bowl. Food cost runs ₹119, labour and rent take roughly ₹150 between them, and the dish clears about ₹71 on a dine-in cover. Put the same bowl on a delivery platform at a 30% commission and ₹102 comes off the top, so the order finishes underwater. This is an illustration to show where the maths turns, not a benchmark for your outlet.
Run your own version of that sum on your three best-selling dishes. A break-even calculator does the arithmetic for you. Knowing your profit margin per dish is what separates an owner who adjusts from one who finds out in March.
Two formats worth reading separately, because they fail differently: cloud kitchens and small ice cream parlours.
Conclusion
The failure rate you were told is wrong, and the one you actually want does not exist for India. What survives scrutiny is 26% in year one and 59% over three years, from American data collected in the late 1990s.
Neither number tells you anything about your outlet. Your rent, your food cost, your churn and your commission do.
Across the kitchens we work with, the ones that make it past year three are rarely the ones with the best food. They are the ones whose owners can tell you what sold yesterday and what it cost them. A restaurant POS is how most of them know.
FAQs
About 26% in the Columbus study, and 17% in the larger Bureau of Labor Statistics census. Both are credible, and they differ because they count differently. One follows a single year’s intake of new restaurants until they close. The other counts closures across a whole database in one year. Always ask which kind of measure you are looking at before setting one against another.
No. It came from a 2003 American Express commercial, and no study has ever produced it. The reason it survives is that it flatters everyone: it makes survivors look heroic and lenders look careful. Treat anyone still quoting it as someone who has not checked.
There is no reliable five-year figure anywhere, which is why the “80% by year five” line has no study behind it. The Columbus data stops at three years, at 59%. If you need a planning assumption, use your own lease term and your break-even month instead of somebody else’s percentage.
Because nothing tracks closures. That gap also explains why so many invented Indian percentages circulate: there is no official figure to contradict them. If you see one quoted, ask for the sample size and the years covered. Neither will be there.
Running out of working capital, and it is almost never one dramatic mistake. Rent, food cost and churn grind together over two or three quarters. The day-end report is the earliest place most owners could have seen it coming, usually months before the bank balance said so.
