Restaurant Site Selection: 15 Data Points That Predict Success

A first-time owner signs a ten-year lease on a beautiful corner space. Exposed brick, great light, a price per square foot that feels like a steal.

Eleven months later, the doors are chained shut.

Here's the part that should keep you up at night: the food was good. The reviews, the few there were, averaged 4.5 stars. The owner did almost everything right. The one thing they got wrong was the only thing they couldn't undo — the address. The "steal" rent was cheap because 40,000 cars a day drove past at 45 mph with no easy way to turn in, and the nearest neighbor with foot traffic was a tire shop that closed at 5 p.m.

Industry research consistently pegs the multi-year failure rate for new restaurants around 60%, and a startlingly large share of those closures trace back not to bad food or bad service, but to a location that was wrong from day one. Get the site right and your survival odds flip dramatically in the other direction.

But here's the thing: site selection isn't luck, and it isn't a gut feeling about "good energy." It's a scoring exercise. Fifteen measurable data points, each one a number you can pull before you ever talk to a landlord. Walk through all fifteen and you stop choosing locations the way amateurs do — by how the empty room makes you feel — and start choosing them the way operators with 15 and 19 stores do: by what the data says the room will do.

Let's go through every one.

Data Points 1-4: Who's Actually Driving and Walking Past

Traffic is the raw fuel of a restaurant. But "lots of traffic" is a rookie metric. The operators who win look at which traffic, going which direction, at which time of day.

1. Vehicle traffic count (VPD). Your state DOT publishes average vehicles-per-day for nearly every road. A drive-thru or quick-service concept generally wants 20,000-40,000+ VPD on the adjacent street. A destination full-service spot can survive on far less if the other numbers are strong. The mistake is treating a big VPD number as automatically good — a six-lane highway where nobody can turn in is worse than a 12,000-VPD main street where they can.

2. Pedestrian count. In dense urban and mall locations, footfall matters more than cars. Stand on the sidewalk at lunch, at dinner, and on a Saturday and physically count people for 15 minutes each. Multiply out. You'll learn more in 45 minutes than from any brochure.

3. Daypart alignment. Here's where it gets worse for a lot of locations: traffic that peaks at the wrong time is useless to you. An office-park lunch concept needs weekday daytime traffic. A dinner-and-drinks concept needs evening and weekend residential traffic. A site can have huge counts that all happen during hours you're either closed or slow.

4. Direction and "going-home" side. Commuters impulse-buy dinner on the way home, not on the way to work. The side of the street that catches the evening homebound flow can outperform the opposite curb by double digits, even though they share the same VPD number.

Data Points 5-8: The People Within Reach

Traffic gets people past your door. Demographics decide whether those people are your people. Define a trade area — typically a 1-mile ring for quick-service, 3-5 miles for full-service and destination dining — and pull these numbers from census data or a site-selection data service.

5. Population density and trade-area count. How many people live and work inside your realistic drive time? Thin density means you're fighting for a tiny pool from day one.

6. Household income. Match income to your average check. A $14 fast-casual bowl and a $60 steakhouse want very different income bands within reach. Aiming a premium concept at a value trade area — or vice versa — is one of the most common and most fatal mismatches.

7. Age and life stage. Bubble tea and late-night ramen index toward younger, student-heavy areas. Family-style and brunch concepts want households with kids. The census tract tells you which one you're walking into.

8. Daytime vs. residential population. A downtown block might have 50,000 daytime workers and 800 residents — perfect for lunch, brutal for dinner. A suburb is the reverse. Know which population you're feeding and when.

Data Points 9-11: Can They See It, Reach It, and Park?

9. Visibility. A restaurant invisible from the road pays for that invisibility in marketing dollars forever. Score the site on how many seconds of clear sightline a passing driver gets, and whether your signage will actually be readable. A monument sign on a busy corner is worth real money; a unit buried at the back of a strip with no road frontage is a permanent handicap.

10. Ingress and egress. Easy in, easy out. A median that forces a U-turn, a left turn across four lanes with no light, or a parking lot you can only enter from a side street will quietly suppress sales every single day. Drive into the site yourself during rush hour and feel the friction.

11. Parking ratio. Count the spaces and divide by your seat count. Full-service restaurants generally need roughly 10-15 spaces per 1,000 sq ft, and customers who circle the block twice often just keep driving. In shared centers, check whether the anchor tenant's customers will eat your spaces at your peak.

Data Points 12-13: The Neighbors and the Competition

12. Co-tenancy. This is the data point amateurs ignore and pros obsess over. The businesses around you decide who's already walking by. A grocery anchor, a gym, a packed coffee shop, a cinema — each generates a predictable flow of people at predictable times. The art is borrowing traffic from neighbors whose customers overlap with yours. The trap is a center full of vacancies or businesses whose peak hours are the mirror image of yours, leaving you to manufacture 100% of your own demand.

13. Competition density. Competition isn't automatically bad — a "restaurant row" can lift everyone by becoming a dining destination. What you're scoring is saturation of your specific concept. Being the fourth taco shop on one block is a problem; being the only ramen shop near a college with three burger joints is an opportunity. Map every competitor in your trade area and ask whether the market is underserved or already full for what you sell.

Data Points 14-15: The Money in the Fine Print

14. Occupancy cost as a percentage of projected sales. Rent in dollars per square foot is meaningless on its own. What matters is total occupancy cost — base rent plus CAM (common area maintenance), property taxes, and insurance pass-throughs — as a share of realistic sales. Healthy restaurants keep that between 6% and 10%. Quick-service aims low because margins are thin; full-service can stretch toward 10-12%. If a space only pencils out at sales numbers you have to squint to believe, the rent is too high — full stop. Run the math both ways before you fall in love. Our break-even calculator will show you the exact monthly sales an address demands before it earns you a dollar.

15. Lease terms and exit options. The number on the rent line is only half the lease. Scrutinize the length, the annual escalations, who pays for build-out (the tenant-improvement allowance), the personal guarantee, the exclusivity clause that keeps a competitor out of your own center, and — critically — the assignment and sublease rights that let you exit or sell if things go sideways. A favorable lease on an average site can outperform a brutal lease on a great one.

Turn Your Own Data Into Your Sharpest Site-Selection Tool

If you're opening your first location, the fifteen points above are your playbook. But if you already run one store, you're sitting on something most consultants would charge five figures to estimate: proof of where your customers actually come from.

This is exactly how multi-location operators scale without rolling the dice. Take T. Jin China Diner — 15 stores, 75 terminals — managing every location's performance in real time from one dashboard. When a brand like that scouts store number 16, it doesn't start from a blank census map. It starts from its own loyalty and CRM data: the home ZIP codes of repeat guests, the drive distances people tolerate, the dayparts that actually pay the rent. Crafty Crab Seafood — 19 stores, 152 terminals — uses the same approach, then keeps every new opening on-brand with one-click menu sync so a new address is selling the proven menu on day one.

Here's how the right POS quietly de-risks your next lease:

• Loyalty & membership records map your true trade area. When your points-and-rewards program shows that 70% of repeat customers live within 4 miles and a meaningful slice drive in from one specific suburb, that suburb is your next site search — backed by data, not a hunch.
• Gift card and e-gift card redemption patterns reveal how far your brand travels. Cards bought at one store and redeemed across town tell you the name carries beyond its immediate neighborhood — a green light for a second address in that direction.
• Daypart sales mix at the checkout tells you whether you're a lunch brand or a dinner brand, which directly decides whether your next site needs office traffic or residential traffic.
• One real-time dashboard across every store lets you compare a new location's ramp against your proven units from the first week, so a soft opening becomes a signal you can act on instead of a number you discover at tax time.

This is where being on a platform built for restaurants pays off long after the lease is signed. KwickOS runs as a hybrid local + cloud system, so the POS keeps ringing sales at 1ms local speed even if the internet at a brand-new address isn't sorted out yet — no small thing during a chaotic opening week. It's processor-agnostic, so you keep 100% of your processing revenue and negotiate your own rates at every location instead of being locked into one vendor's markup. Fingerprint 1:N employee verification travels with you to the new store on day one, and the whole system speaks English, Chinese, and Spanish out of the box — useful when your new trade area's labor pool doesn't match your last one.

And when your shiny new location is humming, the same platform runs your multi-location menu sync and lets you launch delivery through KwickDriver at a flat $2 + $6.99 per 5 miles instead of bleeding 15-25% in commission to third-party apps on every order that new address generates.

The Scoring Sheet: Don't Sign Without It

Before you sign anything, build a simple one-page scorecard. List all 15 data points down the left. Score each candidate site 1-5. Weight the ones that matter most for your concept — traffic and demographics for QSR, co-tenancy and visibility for destination dining — and total it up.

The discipline of forcing a number onto every factor does something powerful: it stops the beautiful exposed-brick room from hypnotizing you past the fact that nobody can turn left into the parking lot. The room you fall in love with and the room that pays your mortgage are not always the same room. The scorecard keeps them honest.

One more lens worth applying: read your candidate sites alongside your revenue-per-square-foot targets. A smaller, more expensive space that generates higher revenue per square foot can easily beat a cheap, cavernous one you'll struggle to fill — and that single metric often settles the debate between two finalists.

The lease is the one decision you can't walk back. Spend the extra two weeks. Pull every number. Score every site. Because the difference between the 60% that close and the survivors usually isn't the recipe — it's the address, and you only get to choose it once.

Open Smarter at Every Address

KwickOS gives multi-location operators one real-time dashboard, processor freedom, offline-proof checkout, and loyalty data that turns your next site search into a pattern match. See how the platform scales with you.

Explore KwickOS for Restaurants

Frequently Asked Questions