How to Choose a Restaurant Location: The 10% Rent Rule and the Checklist Chains Use

The food business has an uncomfortable rule: the kitchen decides whether customers come back, but the address decides whether they show up at all. Ohio State's landmark restaurant-failure research identified location among the top reasons restaurants close (Parsa et al., 2005), and our own reading of the data in the restaurant failure rate guide shows about 17% of independents close in year one, with poor site selection one of the three root causes that show up in study after study.

This guide translates the site-selection playbook that US chains run with data teams into steps an independent owner can execute with a notebook, three days, and one spreadsheet: the rent math, the crowd analysis, the chain tactics worth copying, the franchise trap, and the four locations to never lease.

Why is location the number one restaurant killer?

Because rent is the one big cost you cannot manage after signing. Food cost drifts and you rewrite the menu. Labor spikes and you rework the schedule. But occupancy is fixed for the life of a five or ten year lease, and restaurant net margins run 3% to 5% in a normal year, so a location that overcharges by even three points of sales consumes most of the profit line before the first order is cooked.

The failure data is blunt about it. Restaurants in poor locations fail at rates far above the industry baseline regardless of food quality, and the highest failure clusters sit in the flashiest, highest-rent districts (Parsa et al., 2005). The expensive corner with the beautiful storefront is often the trap, not the prize: maximum occupancy cost, maximum competition, and a customer base that may be passing through rather than buying.

That is why the operating rule of this entire guide is a mindset shift. On the day you pick a site you are not a chef, you are an investor placing capital in a piece of real estate, and the only question that matters is whether the money already moving through that trade area can cover your rent at a healthy ratio.

How do McDonald's, Starbucks, and Domino's choose locations?

They remove emotion from the decision and buy data instead. Harry Sonneborn, McDonald's first president, described the company to investors as a real estate business rather than a burger business (John F. Love, "McDonald's: Behind the Arches", 1986), and the site-selection tactics that grew out of that mindset are worth copying at any scale:

• The commute side of the street. Starbucks site selectors famously favor the morning, going-to-work side of the road. US drivers resist turning left across traffic for a coffee, so the same corner can be a winner on one side of the street and a graveyard on the other. Roughly 70% of major fast-food volume moves through the drive-thru lane (QSR industry reporting), which makes ingress and egress, curb cuts, and turn friction part of the product.
• Anchor and co-tenancy psychology. Big brands rarely try to generate traffic from scratch. They attach themselves to crowds that already exist: grocery anchors, big-box centers, restaurant rows with two or three proven operators. A customer near an anchor store has already switched into spending mode, and leasing brokers price that mindset as co-tenancy.
• Delivery-radius math. Domino's picks sites by counting households and drive times inside the delivery zone, not by admiring the dining room. Since 2018 it has described a strategy it calls fortressing (Domino's investor communications): deliberately adding stores closer together so every address sits minutes from an oven. The lesson for an independent: if delivery or pickup is your model, household density inside a 10 minute drive matters more than frontage.

None of this requires a data science team. It requires asking the chain questions about your own candidate site: which side of the commute is it on, what anchor already gathers my customer, and how many households can reach me in 10 minutes?

What is the difference between a moving crowd and a buying crowd?

A moving crowd is on its way somewhere else; a buying crowd came to spend. Confusing the two is the classic first-timer mistake, because raw foot traffic photographs beautifully and pays nothing.

Two details sharpen the table. First, purchasing power beats headcount: a college strip delivers enormous traffic that caps your average ticket, so a premium concept starves amid a full sidewalk. Second, hurry beats hunger: a commuter rushing for a train passes the best fine-casual concept in the state, which is why transit locations belong to formats measured in seconds.

How do you run a manual footfall count?

Sit at the site for 3 consecutive days, including one weekend day, and watch it in the morning, at lunch, and in the evening. That is the whole method, and it is the cheapest underwriting you will ever do on a six-figure decision:

1. Count, then classify. Not just how many people, but who: families or teenagers, office workers or tourists, buyers carrying bags or walkers passing through.

2. Watch the money. What are people spending at the neighboring businesses? A line at the $6 smoothie window tells you the trade area supports impulse spend; an empty premium salad shop next door is a verdict.

3. Time the friction. Sit in your car and try to enter the lot at 12:30pm and 6:30pm. Count how long the left turn takes, whether parking requires circling, and where the nearest signalized intersection is.

4. Layer the free data. US Census QuickFacts gives household income and density for the surrounding area, and Google Maps Popular Times shows when the neighbors actually get busy. Chains pay data vendors for polished versions of the same signals.

Three days feels slow next to a broker's glossy one-pager. It is also how you catch the site that looks alive at Saturday dinner and dead the other 18 shifts of the week.

What is the 10 percent rent rule?

Your total occupancy cost (base rent plus NNN charges: property taxes, insurance, and common-area maintenance) should not exceed 10% of your expected sales, and the healthiest US independents hold it near 6% to 8% (industry benchmarks from NRA operations data and the major restaurant-platform guides). Above 10%, the site has to outperform just to break even; near 15%, the lease is eating the profit line whole.

The discipline is to run the math backward before you fall in love:

Then apply the mirror test: stand at the site and ask whether this address, with this crowd, can hand you $2,700 across the counter every single day, including the rainy Tuesdays. If the honest answer is no, the location is wrong no matter how beautiful it is, because rent hollows a restaurant out from the inside. Run your own numbers through the break-even calculator, and see the 30/30/30 rule for where occupancy sits inside the full cost structure.

Two US-specific notes. Read the NNN line as carefully as the base rent, because taxes and CAM commonly add 20% to 30% on top and they escalate. And if your concept is prep-heavy, remember the rent arbitrage documented in our central kitchen guide: US retail space asks an average of $24.59 per square foot while industrial space runs $10 to $11 (CBRE and Cushman & Wakefield, Q1 2026), so square footage that cooks but never seats a guest may not belong on retail rent at all.

How do franchise site approvals trap first-time owners?

A franchise site approval is not underwriting, and treating it as one is an expensive mistake. The franchisor's economics are collected up front: an initial franchise fee that commonly runs $20,000 to $50,000, margins on the build-out and equipment package, and royalties that begin the day you open. Whether your specific address thrives affects them far less than it affects you, and most Franchise Disclosure Documents say exactly that in writing: site approval is not a representation that the location will succeed.

So use the FDD the way a lender would, under the FTC's Franchise Rule:

• Item 19 contains whatever financial performance the franchisor is willing to put on paper. If it is empty or vague, price that silence.
• Item 20 lists openings, closures, and transfers by year. Closure clusters are the site-selection track record the sales deck leaves out.
• Your own count still applies. The 3-day footfall count and the 10% rent test do not get waived because a brand approved the address. If the numbers fail, the logo on the door will not save them.

The clean rule: never advance the deposit on a location because the franchisor blessed it. Their approval is a marketing step in their sales process, not a survey of yours.

Which locations should you never lease?

Four site types kill restaurants reliably enough that the discount is never worth it:

1. The spot that killed three concepts in five years. Every market has one: a great-looking space where a taqueria, a wing bar, and a cafe all opened and closed. The overconfident read is that all three operators were fools. The statistical read is that the site has an invisible defect (a broken layout, missing hood or grease infrastructure, hostile ingress, a parking trick, a safety perception after dark) and it will process your savings exactly the way it processed theirs.

2. No parking where guests drive. In most of suburban and small-metro America, dinner arrives by car. If a family of four cannot park within a short walk, they do not become a walk-in, they become a customer of the place 500 yards down the road with a lot. Expecting demand to overcome parking friction is betting against every convenience trend of the last decade.

3. The median out front. A divider in the road makes your site right-in, right-out only: you serve one direction of traffic and lose most of the other, because almost nobody executes a U-turn for lunch. Corner sites at signalized intersections exist precisely to solve this; a mid-block site behind a median only looks equivalent on the brochure.

4. Zero visibility. At 35 mph a driver covers about 150 feet in 3 seconds; if your sign cannot be found and read in that window, drive-by traffic does not exist for you. Basements, second floors, and storefronts hidden behind trees or monument signs all fail this test. What gets seen gets sold, and today the second half of visibility is digital: a site that is hard to describe is also hard to find on a map app, which is why local SEO now belongs in the site walk-through.

Does a strong location matter if most of your orders are online?

Yes, but online volume changes what you are paying rent for. A location sets two ceilings: the walk-in ceiling (frontage, visibility, parking) and the cost ceiling (occupancy as a share of every dollar). A commission-free direct ordering channel raises the first ceiling by extending your trade area beyond the people who physically pass the door: the customer three neighborhoods away who orders pickup twice a month never sees your frontage, and delivery-first concepts push this to the limit by trading visibility away entirely for industrial-cheap rent, then spending the difference on customer acquisition.

The rent math and the channel math also interact. A B-plus location at 7% occupancy with a strong direct channel routinely out-earns an A location at 13%, because the direct orders arrive without the 15% to 30% marketplace toll documented in our delivery app commission tracker; on a flat-fee platform like DirectOrders ($249 per month, zero commission), added volume from the wider trade area lands at card-processing cost. The playbook for building that channel is in how to get more direct orders. None of it substitutes for the 10% rule: it just means the modern site decision weighs a delivery polygon alongside the sidewalk.

The investor's checklist before you sign

Emotions folded and set aside, this is the whole method in eight lines:

1. Classify the crowd: buying, moving, or low-budget, from 3 days of your own counting.

2. Confirm the commute side, the turn friction, and the parking at peak, from your own car.

3. Count households and incomes in the 10 minute trade area (Census QuickFacts, Popular Times).

4. Run the 10% test on total occupancy, NNN included, against honest daily sales.

5. Pass the mirror test: can this address produce that number on a rainy Tuesday?

6. Pull the site's history: anything that closed here in 5 years, and why.

7. Read Items 19 and 20 before believing any franchise site approval.

8. Price the online channel: the delivery polygon and a direct ordering page widen a good location and cannot rescue a bad one.

Where the money already circulates, where the rent fits inside 10 points, and where the customer spends zero effort to reach you: set up the game there.

Sources

• Parsa, Self, Njite, King: Why Restaurants Fail (Cornell Hotel and Restaurant Administration Quarterly, 2005)
• John F. Love, "McDonald's: Behind the Arches" (Bantam, 1986): the Sonneborn real estate quote
• Howard Schultz, "Pour Your Heart Into It" (Hyperion, 1997): Starbucks site-selection discipline
• Domino's investor relations: fortressing strategy communications, 2018 to 2024
• FTC Franchise Rule: FDD requirements, Items 19 and 20
• US Census QuickFacts: free trade-area demographics
• CBRE US retail asking rents and Cushman & Wakefield industrial rents, Q1 2026, as compiled in our central kitchen guide
• National Restaurant Association operations data and 2026 State of the Industry: occupancy and cost benchmarks