How to Start a Restaurant
Opening a restaurant works best in this order: settle the concept, build a costed menu, work out the covers and average spend you need to break even, and only then go looking at sites. Most first-timers fall in love with a property first and reverse-engineer a business to fit it, which is how you end up with rent that no realistic level of trade can carry. Anyone quoting you a single figure for what it costs to open is guessing; the number depends almost entirely on whether you take on a fitted site or a shell.
Figures verified July 2026
Pin the concept down to a sentence
Not a mood board. A sentence that tells you who's coming in, what they're spending, and at what time of day. Everything downstream is decided by that: a 40-seat neighbourhood bistro doing two sittings on a Friday is a different business from an all-day cafe with a 90-second queue at eight in the morning, and they want different sites, different kitchens and different staff. If you can't say it in a sentence, you'll discover the ambiguity later, and it'll be expensive.
Work out what you need to take, and be honest about it
Now you can build the number that matters: covers per service and average spend needed to cover everything. Use the break-even calculator with your real fixed costs, and sanity-check the result against your seat count and how many times you can realistically turn a table. If break-even needs three sittings on a Tuesday, the model is wrong and no amount of good cooking fixes it. Check your food and labour assumptions against the full-service restaurant benchmarks while you're there.
Only now go looking at sites
With a break-even number in hand you can walk into a viewing and know within a minute whether the rent works. Ask what the previous tenant traded, what's included in the fit-out, and what the service charge and rates actually come to, because headline rent is rarely the whole bill. A fitted site with working extraction costs a fraction of a shell, and extraction is the thing that most often turns a cheap unit into an expensive one.
Licences, registration and inspections
Food business registration, a food safety management system, planning permission for any change of use, an alcohol licence if you need one, and in some places a separate late-opening permission. Timelines vary a lot by authority and several of these run in sequence rather than in parallel, so start early and ask your local authority directly what they need. Getting this wrong doesn't cost money so much as months.
Staffing and what an hour really costs
Build the rota from your service pattern, not from a headcount you picked. Then find out what those hours actually cost: statutory floors are only the start, with employer taxes, holiday pay and pension contributions sitting on top. Our labour costs by region pages carry the current rates with the government source against each one. Once you have a rota and a wage bill, the labour cost percentage calculator tells you whether it's viable, and the employee schedule maker will get the first week's rota onto paper.
Check food and labour together before you commit
Food cost and labour cost trade against each other, so judging them separately hides trouble. Put both into the prime cost calculator. If the combined figure leaves nothing for rent, you've found the problem while it's still cheap to fix. This is the check that most business plans skip and most failed restaurants needed.
Decide how bookings and waiting will work
Before opening week, settle how people get a table. A booking system, a walk-in policy, and something for the inevitable Friday when twelve people are standing by the door. Write your no-show and cancellation rules down with the reservation policy generator, print a waitlist sign-in sheet for the host stand, and put a QR code on the menu or the door. None of this is expensive and all of it is worse to improvise in week one.
Open softer than you want to
A soft opening exists so your kitchen can find its timings with a room you can afford to disappoint. Invite people you know, run a short menu, and watch where the service breaks. Whatever your ticket times look like on a friends-and-family night, they will be worse on a full Saturday, and that gap is the thing to fix before you invite a reviewer.
What to budget for
We haven't put figures against these, because the honest range is too wide to be useful and nobody publishes a dependable average. What we can do is make sure nothing on the list catches you out.
| Cost | Worth knowing |
|---|---|
| Lease deposit and first rent | Often several months up front, plus a personal guarantee. |
| Fit-out and extraction | The single biggest swing between a fitted unit and a shell. |
| Kitchen equipment | Buy used where reliability allows, new where a failure closes you. |
| Furniture and front of house | Cheaper to underspend here than in the kitchen. |
| Licences, registration and professional fees | Solicitor, accountant, planning, alcohol licence. |
| Opening stock | Food, drink, packaging, cleaning, small wares. |
| EPOS, booking and payment systems | Setup fees plus ongoing monthly cost. |
| Insurance | Public liability, employer liability, buildings and contents, business interruption. |
| Pre-opening wages | Staff are paid to train before a single cover is served. |
| Working capital | The one everyone underestimates. Budget to trade at a loss for months, not weeks. |
Where this usually goes wrong
Signing a lease before the numbers exist
Rent is the one cost you cannot renegotiate once you've committed, and it's fixed whether you're full or empty. If you take one thing from this page: build the break-even model first, then look at sites.
Running out of cash while trading fine
Plenty of restaurants close while doing decent covers, because opening costs ate the buffer and there was nothing left for a quiet January. Profit and cash aren't the same thing, and the gap between them is what kills first-year businesses.
A menu that's too long
Long menus mean more stock lines, more waste, slower tickets and a kitchen that never gets good at anything. Almost every experienced operator cuts the opening menu within six months. Start where you'd end up.
Treating the waiting area as a design problem
Walk-ins who leave before being seated leave no trace in your figures, so a bad Friday just feels like a slow one. On a full Friday this is real money, and it's the easiest revenue in the building to recover.
What to do about the waiting
Once you're open, the cheapest sales in the building sit in the gap between a table being cleared and the next party sitting down, and in the walk-ins who give up waiting. Both are throughput problems rather than marketing ones, and both get worse on exactly the nights you can least afford it.
Most operators treat a queue as pure cost and try to shorten it. It is also the one stretch of time when a customer is holding a phone specifically because of you, which makes it the cheapest attention you will ever get. Sorting out a digital waitlist and guest paging system for restaurants before you open handles both sides: people stop drifting off, and the screen they are watching can carry an offer. There is more on that idea in monetising customer wait times and upselling while customers wait.
Run your own numbers
Free tools you'll want on day one
Once you're trading, compare your numbers against the published benchmarks for this kind of business, and check what staff cost where you are on our labour costs pages. Everything sits together under industry cost benchmarks.
The wait is the only ad space your guests actually watch
Guests waiting for a table sit on the tracker checking their place in the queue. That screen is yours: put the cocktail list, the specials or a dessert on it while they are already looking. They get told the moment the table is ready, and the wait pays for itself instead of costing you walkaways.
Frequently asked questions
How much does it cost to open a restaurant?
There's no honest single figure, and anyone quoting one is guessing. The range is enormous depending on whether you take a fitted site or a shell, since extraction and kitchen fit-out dominate everything else. Build your own number from the budget lines on this page, then pressure-test it with the break-even calculator.
What should I do first when starting a restaurant?
Define the concept in a sentence, then cost the menu, then build a break-even model. Site hunting comes fourth. Doing it in that order means you can judge a property in a minute instead of talking yourself into it.
Do I need a business plan to open a restaurant?
You need one if you're raising money, and you need the underlying numbers regardless. The useful core is a costed menu, a break-even model, a staffing plan with real wage costs, and a cash flow that assumes a slow start.
What licences does a restaurant need?
Typically food business registration, a documented food safety management system, planning permission where the use is changing, and an alcohol licence if you serve it. Requirements and timelines vary by authority, so ask yours directly rather than working from a generic checklist.
How long before a new restaurant makes money?
Longer than most plans assume. Budget working capital to trade through a slow opening period and a quiet season after it, and treat any earlier profitability as a bonus rather than the plan.
What percentage of restaurants fail?
Widely quoted failure rates circulate without much behind them, so we won't repeat one. The recurring causes are more useful than the statistic: rent that trade can't carry, running out of cash, and a menu too broad to execute well.
Starting something else
- How to Start a Takeaway or Takeout BusinessCheaper to open than a restaurant, and the channel mix decides whether it works.
- How to Start a Food TruckNo rent, but pitch fees, commissary hire and a revenue line that swings hard.
- How to Start a Coffee ShopA small ticket, a ninety-minute peak, and a site decision you cannot undo.
This guide covers general practice and isn't legal, financial or licensing advice. Requirements differ by country and by local authority, so confirm anything to do with permits, planning or employment law with yours before you commit money to it.