The 9 Key KPIs for Independent Restaurants in 2027

👁 0 views📖 2,784 words⏱ 13 min read📅 Published Jun 3, 2026

TL;DR
— Independent Restaurants in 2027: The nine KPIs that separate surviving independents from the 9,500+ locations that closed in 2025 are food cost % (target 28-32%), labor cost % (target 28-32%), prime cost (hard ceiling 65%), RevPASH ($7-$12 casual, $15-$30 fine), average check, table turn time (target 3 turns/lunch, 2 turns/dinner), customer count, third-party delivery mix (cap at 15-20% of sales), and repeat-visit rate (target 35%+)**.
Generic SaaS dashboards do not apply — restaurants run on perishable inventory, hourly labor, and ticket-by-ticket margin that can swing 8 points based on a single bad shift.

Why Independent Restaurants Report Differently

Independent restaurants are not SaaS companies, and the KPIs that work for a B2B software business will bankrupt a 90-seat bistro. Three structural realities force a different reporting stack.

First, perishable inventory. A SaaS company can sit on unsold seats forever; a restaurant throws out $0.04-$0.07 of every dollar in food waste if the inventory system is sloppy. That single fact makes food cost % the most-watched line in the building — not MRR, not NRR, not ARR.

The National Restaurant Association's 2026 State of the Restaurant Industry report flagged that more than 90% of operators feel direct margin pressure from food inflation alone.

Second, hourly labor with built-in overtime traps. A salaried SaaS engineer is a fixed cost; a line cook at $18-$24/hour in 2027 is a variable cost that can blow up the P&L in a single Saturday night if the manager over-schedules. Full-service labor now runs at a median 36.5% of sales per Toast's 2026 industry payroll study, with profitable operators holding it to 34.2%.

That 2.3-point delta is often the entire difference between a profitable year and bankruptcy.

Third, seat-time as the actual unit of inventory. Software firms measure logo count; restaurants measure revenue per available seat hour (RevPASH) because a 60-seat dining room open 8 hours a day has exactly 480 seat-hours of inventory every day, and they all expire at midnight.

Cornell's School of Hotel Administration has anchored RevPASH as the operational equivalent of RevPAR for hotels since the late 1990s, and it remains the single best composite KPI a 2027 independent can run.

flowchart TD A[Food Cost %] --> P[Prime Cost] B[Labor Cost %] --> P P --> M[Restaurant-Level Margin] C[Average Check] --> R[RevPASH] D[Table Turn Time] --> R E[Customer Count] --> R R --> M F[3rd-Party Delivery %] --> P F --> M G[Repeat-Visit Rate] --> E G --> H[Marketing CAC Avoided] H --> M M --> S[Operator Take-Home + Reinvestment]

The independent sector contracted 2.3% in 2025 per Datassential / Restaurant Business Online — full-service indies down 2.6%, limited-service indies down 1.8% — which means in 2027 the discipline of measuring these nine numbers weekly is not a "best practice." It is the survival kit.

The 9 KPIs, In Depth

1. Food Cost Percentage

Definition. Cost of food sold divided by food revenue, measured weekly on a true inventory basis (beginning inventory + purchases − ending inventory).

Formula. Food Cost % = (Beginning Inventory + Food Purchases − Ending Inventory) / Food Sales

Benchmark. 28-32% is the 2027 target band for independents per Restaurant365, 7shifts, and Toast benchmark studies. Fine dining trends higher (32-38%) because of premium proteins; pizza/pasta concepts trend lower (22-28%).

Named operator example. Sweetgreen (publicly disclosed in its 10-K) ran food, beverage, and packaging at 27.4% of revenue in 2024 — proof that a vertically-integrated salad concept can press well below the band. Independents rarely match it, but Honest Mary's in Austin has publicly cited 28-30% on a fast-casual grain bowl menu.

Failure mode. Operators count purchases as food cost and skip the inventory bookend. They miss spoilage and over-portioning until quarter-end, by which point two to four points of margin have evaporated.

2. Labor Cost Percentage

Definition. Total labor (wages + payroll taxes + benefits + workers' comp) divided by net sales.

Formula. Labor Cost % = (Wages + Payroll Taxes + Benefits + Workers' Comp) / Net Sales

Benchmark. 28-32% for healthy full-service independents; 20-25% for quick service. Toast's 2026 payroll study put the full-service median at 36.5% with profitable operators at 34.2% — a clear sign most independents are over-staffed or under-priced.

Named operator example. Texas Roadhouse (NASDAQ: TXRH) disclosed labor of 33.4% of sales in its 2024 10-K despite operating in a tip-credit-favorable state mix; an independent steakhouse without scale leverage typically sits 2-4 points worse.

Failure mode. Scheduling by gut, not by forecasted covers. A manager who staffs the same Tuesday-night line-up as Saturday is silently giving away 3-5 points of labor.

3. Prime Cost

Definition. Food cost + labor cost, expressed as a percentage of sales. The single most important composite number in the building.

Formula. Prime Cost % = (Total Food Cost + Total Labor Cost) / Total Sales

Benchmark. 55-65% is the universal target band per Baker Tilly, Restaurant365, and The Restaurant Group. Above 65% there is essentially nothing left for rent, utilities, and profit. Full-service should target 60-65%; QSR 55-60%; fine dining sometimes runs 66-68% and survives only because average checks are high.

Named operator example. Chipotle Mexican Grill (NYSE: CMG) ran prime cost at roughly 55% in its 2024 10-K (food + labor combined), reflecting the public-company scale advantage. Shake Shack sits closer to 60-62%.

Failure mode. Operators look at food and labor separately, never sum them, and miss that the business is structurally 66-68% — a death-spiral configuration that no marketing campaign can fix.

4. RevPASH (Revenue Per Available Seat Hour)

Definition. Total revenue divided by (seats × operating hours). Originated at Cornell's School of Hotel Administration as the restaurant-side analog to RevPAR.

Formula. RevPASH = Total Revenue / (Number of Seats × Operating Hours)

Benchmark. $3-$6 for QSR, $7-$12 for casual, $15-$30+ for fine dining per Boston University's published RevPASH research and 7shifts operator data.

Named operator example. A 90-seat New York City casual bistro open 6 hours of effective service with $12,000 in nightly sales generates RevPASH of $22 — top of the casual band, driven by Manhattan check averages and aggressive seat turnover.

Failure mode. Tracking only nightly revenue. Two restaurants with the same $12K night can have radically different efficiency if one has 60 seats and the other 120.

5. Average Check

Definition. Total food and beverage revenue divided by guest count (covers).

Formula. Average Check = Net Sales / Number of Guests

Benchmark. 2027 norms: casual dining $22-$35, polished casual $40-$60, fine dining $95-$180+. Average checks have risen approximately 4-6% year-over-year since 2023 per Black Box Intelligence menu pricing data.

Named operator example. The Cheesecake Factory disclosed an average check around $28 in its 2024 investor materials; Ruth's Chris Steak House runs $95-$110.

Failure mode. Raising prices without re-engineering the menu. Independents who push the check by 5% without redesigning menu architecture lose 3-4% of covers — a net-zero or net-negative move.

6. Table Turn Time

Definition. Average minutes per cover from seat to settle, used to calculate turns per shift.

Formula. Turns per Shift = Covers Served / Available Seats

Benchmark. 3 turns per lunch and 2 turns per dinner is the family-restaurant industry average per TouchBistro benchmark research; fine dining targets 1.5-2 turns per dinner; QSR effectively runs 6-10 turns through counter service.

Named operator example. Chick-fil-A (private) is widely cited at 2x the drive-thru throughput of the QSR industry average — roughly 120-130 cars/hour at peak, the cleanest throughput case study in the U.S. Industry.

Failure mode. Confusing slow turns with "premium experience." A 110-minute casual lunch turn is not premium; it is a kitchen and FOH coordination failure that destroys RevPASH.

7. Customer Count (Covers)

Definition. Total guests served per shift, day, week, month. The denominator behind almost every other KPI.

Formula. Customer Count = Sum of Guests per Check, by Period

Benchmark. For a 80-100 seat full-service casual independent, healthy weekly cover counts are 1,800-2,800 covers/week; below 1,500 covers/week on that seat count, the unit is structurally unprofitable.

Named operator example. Olive Garden (Darden, NYSE: DRI) discloses same-restaurant traffic quarterly — a 2-3% traffic decline is a recession signal; a 1-2% gain is a winning year.

Failure mode. Operators chase average check with menu hikes while traffic quietly declines. Six months later, top-line is flat but the brand has lost a meaningful share of regulars who never come back.

8. Third-Party Delivery Mix

Definition. Percentage of total sales that flow through DoorDash, Uber Eats, Grubhub, and similar third-party marketplaces, net of their commissions.

Formula. 3PD Mix % = 3PD Gross Sales / Total Gross Sales

Benchmark. Cap at 15-20% of sales is the 2027 consensus from The Independent Restaurant Coalition and Restaurant Business Online. Commission tiers per DoorDash, Uber Eats, and Grubhub: 15% basic / 25% standard / 30% premium placement. Real all-in cost typically reaches 30-40% per order once marketing fees and processing are stacked.

Named operator example. Wingstop (NASDAQ: WING) publicly disclosed digital sales of ~70% in 2024, but the majority through its own first-party app — explicit strategy to keep the third-party take rate below 15% of total sales.

Failure mode. Treating 3PD revenue at the same margin as dine-in. A $45 dine-in ticket at 25% prime cost contribution is wildly different from a $45 DoorDash ticket where the platform takes 30%, leaving the operator with 0-2% margin — sometimes negative once packaging is included.

9. Repeat-Visit Rate

Definition. Share of covers/transactions in a period that come from previously-identified guests (loyalty program, POS account, phone number, card hash).

Formula. Repeat-Visit Rate = Repeat Guest Visits / Total Visits

Benchmark. 30-40% is the healthy band per Restroworks and Qubriux retention studies; below 20% is a red flag. 65-80% of restaurant sales come from regulars, which is why this is non-negotiable. The industry's average overall customer retention is roughly 55% versus a global cross-industry benchmark of 75%.

Named operator example. Starbucks Rewards disclosed roughly 60% of U.S. Company-operated sales through its loyalty program in 2024 — the high-end ceiling that independents will never match, but a directional proof that loyalty depth = top-line stability.

Failure mode. No identity capture at all. Without POS-tied loyalty, an independent has no idea whether the Friday-night dining room is 80% regulars (stable) or 80% one-time visitors (cliff risk).

Real Operators

Sweetgreen (NYSE: SG) — 2024 10-K: food, beverage, packaging at 27.4% of revenue; labor at 27.5%; combined prime cost roughly 55%, near the QSR/fast-casual floor.
Chipotle (NYSE: CMG) — 2024 10-K: food, beverage, packaging at 29.7%; labor at 25.0%; prime cost ~55%. AUV approximately $3.2M per unit.
Texas Roadhouse (NASDAQ: TXRH) — 2024 10-K: food/beverage 33.6%; labor 33.4%; prime cost ~67% — the upper end of viable, offset by $7M+ AUVs.
Cheesecake Factory (NASDAQ: CAKE) — 2024 disclosures: food cost 22-23% of revenue (menu engineering), labor 35-36%; average check ~$28.
The Cheesecake Factory also publishes comparable restaurant sales quarterly — a useful "weather vane" for any U.S. Casual independent.
Honest Mary's (Austin, private) — cited in QSR Magazine at 28-30% food cost on its grain-bowl program.

Failure Modes

No weekly P&L. Operators who close the books monthly are flying blind. By the time the 35-day P&L shows up, the bleeding has already happened. Weekly close is the 2027 minimum.
Counting purchases as food cost. Skipping the inventory count means food cost numbers are off by 2-4 points, almost always understated, until the year-end true-up wipes out reported profit.
Tip-credit confusion. Operators in tip-credit states book labor at the direct wage instead of the effective wage and convince themselves labor is 22% when it is really 30%.
Treating 3PD revenue at dine-in margin. Reporting gross delivery sales without netting 30-40% all-in platform cost inflates revenue and hides the negative-margin orders.
Average-check obsession. Pushing the check at the cost of covers. Black Box Intelligence data repeatedly shows operators who lose 2%+ in traffic rarely recover even with 5%+ menu price increases.
No identity capture. Without a POS-tied loyalty layer (Toast Rewards, Square Loyalty, Thanx), the repeat-visit rate is unmeasurable, which means the single most-predictive KPI is dark.

Reporting Cadence

Daily (every shift close). Sales, covers, average check, labor hours vs. Forecast, voids/comps, DoorDash/Uber Eats daily settlement.
Weekly (Monday a.m.). Food cost % (weekly inventory), labor cost %, prime cost, RevPASH by day-part, 3PD mix %, customer count trend.
Monthly. Full P&L close, repeat-visit rate (loyalty cohort report), menu mix analysis, vendor price changes, table-turn-time distribution.
Quarterly. Menu engineering rebuild, 30-60-90 staffing model reset, lease-and-occupancy ratio review, competitive average-check survey.
Annually. Full operator scorecard against National Restaurant Association and Black Box Intelligence peer benchmarks; reset KPI targets for the coming year.

30 / 60 / 90 Day Implementation

flowchart LR A[Day 0: Audit] --> B[Days 1-30] B --> C[Days 31-60] C --> D[Days 61-90] B -->|Install weekly P&L POS clean-up Inventory bookend| B1[Food + Labor visible weekly] C -->|Add RevPASH 3PD audit Loyalty launch| C1[Throughput + retention visible] D -->|Menu engineering Scheduling model Quarterly scorecard| D1[Prime cost <65% sustained]

Days 1-30 — Visibility. Install weekly P&L with true inventory bookend. Clean up POS categories so food, beverage, and modifiers map correctly. Pull 30 days of DoorDash, Uber Eats, Grubhub statements to compute real take rate.

Days 31-60 — Throughput. Build a RevPASH dashboard by day-part. Audit 3PD: drop the lowest-margin platform tier or renegotiate. Launch a POS-tied loyalty program (Toast Rewards or Square Loyalty) and capture phone/email at every check.

Days 61-90 — Discipline. Run menu engineering — stars, plowhorses, puzzles, dogs. Build a forecast-driven scheduling model (sales-to-labor ratio by half-hour). Publish the first quarterly KPI scorecard to the team. Lock prime cost below 65% as the non-negotiable floor.

FAQ

Q: My food cost is "fine" at 35%. Why does my accountant say I'm losing money? Food cost at 35% is high-end for full-service casual. Combined with full-service labor running 35-37%, prime cost is 70%+ — which mathematically leaves nothing for rent, utilities, insurance, or profit. The accountant is right.

Q: I do $1.8M in sales and feel like I'm working for free. What KPI explains it? Almost always prime cost above 65%. On $1.8M revenue, every point of prime cost is $18,000 of operator take-home. Three excess points = $54,000 gone.

Q: Should I drop DoorDash entirely? Only if 3PD is below 5% of your sales and your dine-in is at capacity. Most independents are better served capping 3PD at 15-20%, renegotiating to the 15% lite tier, and pushing customers to first-party online ordering (Toast, Square, Olo).

Q: How do I measure repeat-visit rate without a loyalty program? You can't, accurately. The minimum-viable substitute is credit-card hashing in modern POS systems (Toast, Square, Lightspeed) which can identify card-level repeats without an enrolled loyalty account. Real number still requires opt-in identity capture.

Q: What's the single most-leveraged KPI to fix first? Prime cost. Every single dollar of prime cost reduction drops to the bottom line. A 2-point improvement on a $2M restaurant is $40,000 of profit that did not exist last quarter.

Sources

National Restaurant Association — 2026 State of the Restaurant Industry report (restaurant.org/research-and-media/research/research-reports/state-of-the-industry).
Restaurant Business Online / Datassential — "The independent restaurant sector shrunk by 2.3% in 2025" (nrn.com).
Toast — Restaurant Payroll Percentage Guide, 2026 (pos.toasttab.com/blog/on-the-line/restaurant-payroll-percentage).
Restaurant365 — How to Calculate Prime Cost in a Restaurant (restaurant365.com/blog/how-to-calculate-prime-cost-in-a-restaurant).
7shifts — Restaurant Prime Cost Guide and RevPASH for Restaurants: Definition, Formula, and Best Practices (7shifts.com/blog).
Cornell University School of Hotel Administration — RevPASH research; Boston University RevPASH portal (bu.edu/revpash).
Baker Tilly — Prime Cost Target Tips advisory note (bakertilly.com/insights/prime-cost-target-tips).
TouchBistro — 10 Essential Restaurant Benchmarks and 21 Restaurant Metrics (touchbistro.com/blog).
Independent Restaurant Coalition — Third-Party Delivery Apps Federal Regulation Brief (independentrestaurantcoalition.com/delivery_apps_2025).
Rezku — Third-Party Delivery Fees in 2026: What DoorDash, Uber Eats & Grubhub Really Cost Restaurants (rezku.com).
Restroworks — Restaurant Customer Retention Statistics (restroworks.com/blog/customer-retention-statistics-restaurants).
Sweetgreen (NYSE: SG), Chipotle (NYSE: CMG), Texas Roadhouse (NASDAQ: TXRH), Cheesecake Factory (NASDAQ: CAKE), Wingstop (NASDAQ: WING) — 2024 Form 10-K filings (sec.gov/edgar).

Keep reading

> **TL;DR — Independent Restaurants in 2027:** The nine KPIs that separate surviving independents from the **9,500+ locations that closed in 2025** are **food cost % (target 28-32%)**, **labor cost % (target 28-32%)**, **prime cost (hard ceiling 65%)**, **RevPASH ($7-$12 casual, $15-$30 fine)**, **average check**, **table turn time (target 3 turns/lunch, 2 turns/dinner)**, **customer count**, **third-party delivery mix (cap at 15-20% of sales)**, and **repeat-visit rate (target 35%+)**. Generic SaaS dashboards do not apply — restaurants run on perishable inventory, hourly labor, and ticket-by-ticket margin that can swing 8 points based on a single bad shift.

## Why Independent Restaurants Report Differently

Independent restaurants are not SaaS companies, and the KPIs that work for a B2B software business will bankrupt a 90-seat bistro. **Three structural realities** force a different reporting stack.

**First, perishable inventory.** A SaaS company can sit on unsold seats forever; a restaurant throws out **$0.04-$0.07 of every dollar in food waste** if the inventory system is sloppy. That single fact makes **food cost %** the most-watched line in the building — not MRR, not NRR, not ARR. The **National Restaurant Association's 2026 State of the Restaurant Industry** report flagged that **more than 90% of operators** feel direct margin pressure from food inflation alone.

**Second, hourly labor with built-in overtime traps.** A salaried SaaS engineer is a fixed cost; a line cook at **$18-$24/hour** in 2027 is a variable cost that can blow up the P&L in a single Saturday night if the manager over-schedules. **Full-service labor now runs at a median 36.5% of sales** per Toast's 2026 industry payroll study, with profitable operators holding it to **34.2%**. That 2.3-point delta is often the entire difference between a profitable year and bankruptcy.

**Third, seat-time as the actual unit of inventory.** Software firms measure logo count; restaurants measure **revenue per available seat hour (RevPASH)** because a 60-seat dining room open 8 hours a day has exactly **480 seat-hours of inventory** every day, and they all expire at midnight. **Cornell's School of Hotel Administration** has anchored RevPASH as the operational equivalent of RevPAR for hotels since the late 1990s, and it remains the single best composite KPI a 2027 independent can run.

```mermaid
flowchart TD
    A[Food Cost %] --> P[Prime Cost]
    B[Labor Cost %] --> P
    P --> M[Restaurant-Level Margin]
    C[Average Check] --> R[RevPASH]
    D[Table Turn Time] --> R
    E[Customer Count] --> R
    R --> M
    F[3rd-Party Delivery %] --> P
    F --> M
    G[Repeat-Visit Rate] --> E
    G --> H[Marketing CAC Avoided]
    H --> M
    M --> S[Operator Take-Home + Reinvestment]
```

The independent sector contracted **2.3% in 2025** per **Datassential / Restaurant Business Online** — full-service indies down **2.6%**, limited-service indies down **1.8%** — which means in 2027 the discipline of measuring these nine numbers weekly is not a "best practice." It is the survival kit.

## The 9 KPIs, In Depth

### 1. Food Cost Percentage

**Definition.** Cost of food sold divided by food revenue, measured weekly on a true inventory basis (beginning inventory + purchases − ending inventory).

**Formula.** `Food Cost % = (Beginning Inventory + Food Purchases − Ending Inventory) / Food Sales`

**Benchmark.** **28-32%** is the 2027 target band for independents per **Restaurant365**, **7shifts**, and **Toast** benchmark studies. Fine dining trends higher (**32-38%**) because of premium proteins; pizza/pasta concepts trend lower (**22-28%**).

**Named operator example.** **Sweetgreen** (publicly disclosed in its 10-K) ran food, beverage, and packaging at **27.4% of revenue** in 2024 — proof that a vertically-integrated salad concept can press well below the band. Independents rarely match it, but **Honest Mary's** in Austin has publicly cited **28-30%** on a fast-casual grain bowl menu.

**Failure mode.** Operators count purchases as food cost and skip the inventory bookend. They miss spoilage and over-portioning until quarter-end, by which point **two to four points of margin** have evaporated.

### 2. Labor Cost Percentage

**Definition.** Total labor (wages + payroll taxes + benefits + workers' comp) divided by net sales.

**Formula.** `Labor Cost % = (Wages + Payroll Taxes + Benefits + Workers' Comp) / Net Sales`

**Benchmark.** **28-32%** for healthy full-service independents; **20-25%** for quick service. **Toast's 2026 payroll study** put the full-service median at **36.5%** with profitable operators at **34.2%** — a clear sign most independents are over-staffed or under-priced.

**Named operator example.** **Texas Roadhouse** (NASDAQ: TXRH) disclosed **labor of 33.4% of sales** in its 2024 10-K despite operating in a tip-credit-favorable state mix; an independent steakhouse without scale leverage typically sits **2-4 points worse**.

**Failure mode.** Scheduling by gut, not by forecasted covers. A manager who staffs the same Tuesday-night line-up as Saturday is silently giving away **3-5 points of labor**.

### 3. Prime Cost

**Definition.** Food cost + labor cost, expressed as a percentage of sales. The single most important composite number in the building.

**Formula.** `Prime Cost % = (Total Food Cost + Total Labor Cost) / Total Sales`

**Benchmark.** **55-65%** is the universal target band per **Baker Tilly**, **Restaurant365**, and **The Restaurant Group**. **Above 65% there is essentially nothing left for rent, utilities, and profit.** Full-service should target **60-65%**; QSR **55-60%**; fine dining sometimes runs **66-68%** and survives only because average checks are high.

**Named operator example.** **Chipotle Mexican Grill** (NYSE: CMG) ran prime cost at roughly **55%** in its 2024 10-K (food + labor combined), reflecting the public-company scale advantage. **Shake Shack** sits closer to **60-62%**.

**Failure mode.** Operators look at food and labor separately, never sum them, and miss that the business is structurally **66-68%** — a death-spiral configuration that no marketing campaign can fix.

### 4. RevPASH (Revenue Per Available Seat Hour)

**Definition.** Total revenue divided by (seats × operating hours). Originated at **Cornell's School of Hotel Administration** as the restaurant-side analog to RevPAR.

**Formula.** `RevPASH = Total Revenue / (Number of Seats × Operating Hours)`

**Benchmark.** **$3-$6** for QSR, **$7-$12** for casual, **$15-$30+** for fine dining per **Boston University's published RevPASH research** and **7shifts** operator data.

**Named operator example.** A 90-seat New York City casual bistro open **6 hours of effective service** with **$12,000 in nightly sales** generates **RevPASH of $22** — top of the casual band, driven by Manhattan check averages and aggressive seat turnover.

**Failure mode.** Tracking only nightly revenue. Two restaurants with the same $12K night can have radically different efficiency if one has 60 seats and the other 120.

### 5. Average Check

**Definition.** Total food and beverage revenue divided by guest count (covers).

**Formula.** `Average Check = Net Sales / Number of Guests`

**Benchmark.** 2027 norms: casual dining **$22-$35**, polished casual **$40-$60**, fine dining **$95-$180+**. Average checks have risen approximately **4-6% year-over-year** since 2023 per **Black Box Intelligence** menu pricing data.

**Named operator example.** **The Cheesecake Factory** disclosed an average check around **$28** in its 2024 investor materials; **Ruth's Chris Steak House** runs **$95-$110**.

**Failure mode.** Raising prices without re-engineering the menu. Independents who push the check by **5%** without redesigning menu architecture lose **3-4% of covers** — a net-zero or net-negative move.

### 6. Table Turn Time

**Definition.** Average minutes per cover from seat to settle, used to calculate turns per shift.

**Formula.** `Turns per Shift = Covers Served / Available Seats`

**Benchmark.** **3 turns per lunch** and **2 turns per dinner** is the **family-restaurant industry average** per **TouchBistro** benchmark research; fine dining targets **1.5-2 turns per dinner**; QSR effectively runs **6-10 turns** through counter service.

**Named operator example.** **Chick-fil-A** (private) is widely cited at **2x the drive-thru throughput of the QSR industry average** — roughly **120-130 cars/hour at peak**, the cleanest throughput case study in the U.S. Industry.

**Failure mode.** Confusing slow turns with "premium experience." A 110-minute casual lunch turn is not premium; it is a kitchen and FOH coordination failure that destroys RevPASH.

### 7. Customer Count (Covers)

**Definition.** Total guests served per shift, day, week, month. The denominator behind almost every other KPI.

**Formula.** `Customer Count = Sum of Guests per Check, by Period`

**Benchmark.** For a **80-100 seat full-service casual independent**, healthy weekly cover counts are **1,800-2,800 covers/week**; below **1,500 covers/week** on that seat count, the unit is structurally unprofitable.

**Named operator example.** **Olive Garden** (Darden, NYSE: DRI) discloses **same-restaurant traffic** quarterly — a **2-3% traffic decline** is a recession signal; a **1-2% gain** is a winning year.

**Failure mode.** Operators chase **average check** with menu hikes while traffic quietly declines. Six months later, top-line is flat but the brand has lost a meaningful share of regulars who never come back.

### 8. Third-Party Delivery Mix

**Definition.** Percentage of total sales that flow through **DoorDash, Uber Eats, Grubhub**, and similar third-party marketplaces, net of their commissions.

**Formula.** `3PD Mix % = 3PD Gross Sales / Total Gross Sales`

**Benchmark.** **Cap at 15-20% of sales** is the 2027 consensus from **The Independent Restaurant Coalition** and **Restaurant Business Online**. Commission tiers per **DoorDash, Uber Eats, and Grubhub**: **15% basic / 25% standard / 30% premium placement**. Real all-in cost typically reaches **30-40% per order** once marketing fees and processing are stacked.

**Named operator example.** **Wingstop** (NASDAQ: WING) publicly disclosed **digital sales of ~70%** in 2024, but the majority through its own first-party app — explicit strategy to keep the **third-party take rate** below **15%** of total sales.

**Failure mode.** Treating 3PD revenue at the same margin as dine-in. A **$45 dine-in ticket** at **25% prime cost contribution** is wildly different from a **$45 DoorDash ticket** where the platform takes **30%**, leaving the operator with **0-2% margin** — sometimes negative once packaging is included.

### 9. Repeat-Visit Rate

**Definition.** Share of covers/transactions in a period that come from previously-identified guests (loyalty program, POS account, phone number, card hash).

**Formula.** `Repeat-Visit Rate = Repeat Guest Visits / Total Visits`

**Benchmark.** **30-40% is the healthy band** per **Restroworks** and **Qubriux** retention studies; below **20%** is a red flag. **65-80% of restaurant sales come from regulars**, which is why this is non-negotiable. The industry's average overall **customer retention is roughly 55%** versus a global cross-industry benchmark of **75%**.

**Named operator example.** **Starbucks Rewards** disclosed **roughly 60% of U.S. Company-operated sales** through its loyalty program in 2024 — the high-end ceiling that independents will never match, but a directional proof that **loyalty depth = top-line stability**.

**Failure mode.** No identity capture at all. Without POS-tied loyalty, an independent has no idea whether the Friday-night dining room is **80% regulars** (stable) or **80% one-time visitors** (cliff risk).

## Real Operators

- **Sweetgreen (NYSE: SG)** — 2024 10-K: food, beverage, packaging at **27.4% of revenue**; labor at **27.5%**; combined prime cost roughly **55%**, near the QSR/fast-casual floor.
- **Chipotle (NYSE: CMG)** — 2024 10-K: food, beverage, packaging at **29.7%**; labor at **25.0%**; prime cost **~55%**. AUV approximately **$3.2M per unit**.
- **Texas Roadhouse (NASDAQ: TXRH)** — 2024 10-K: food/beverage **33.6%**; labor **33.4%**; prime cost **~67%** — the upper end of viable, offset by **$7M+ AUVs**.
- **Cheesecake Factory (NASDAQ: CAKE)** — 2024 disclosures: food cost **22-23%** of revenue (menu engineering), labor **35-36%**; average check **~$28**.
- **The Cheesecake Factory** also publishes **comparable restaurant sales** quarterly — a useful "weather vane" for any U.S. Casual independent.
- **Honest Mary's (Austin, private)** — cited in **QSR Magazine** at **28-30% food cost** on its grain-bowl program.

## Failure Modes

1. **No weekly P&L.** Operators who close the books monthly are flying blind. By the time the **35-day P&L** shows up, the bleeding has already happened. **Weekly close** is the 2027 minimum.
2. **Counting purchases as food cost.** Skipping the inventory count means food cost numbers are off by **2-4 points**, almost always understated, until the year-end true-up wipes out reported profit.
3. **Tip-credit confusion.** Operators in tip-credit states book labor at the **direct wage** instead of the effective wage and convince themselves labor is **22%** when it is really **30%**.
4. **Treating 3PD revenue at dine-in margin.** Reporting **gross** delivery sales without netting **30-40% all-in platform cost** inflates revenue and hides the negative-margin orders.
5. **Average-check obsession.** Pushing the check at the cost of covers. **Black Box Intelligence** data repeatedly shows operators who lose **2%+ in traffic** rarely recover even with **5%+ menu price increases**.
6. **No identity capture.** Without a POS-tied loyalty layer (**Toast Rewards, Square Loyalty, Thanx**), the repeat-visit rate is unmeasurable, which means the single most-predictive KPI is dark.

## Reporting Cadence

- **Daily (every shift close).** Sales, covers, average check, labor hours vs. Forecast, voids/comps, **DoorDash/Uber Eats** daily settlement.
- **Weekly (Monday a.m.).** Food cost % (weekly inventory), labor cost %, prime cost, RevPASH by day-part, 3PD mix %, customer count trend.
- **Monthly.** Full P&L close, repeat-visit rate (loyalty cohort report), menu mix analysis, vendor price changes, table-turn-time distribution.
- **Quarterly.** Menu engineering rebuild, **30-60-90 staffing model** reset, lease-and-occupancy ratio review, competitive average-check survey.
- **Annually.** Full operator scorecard against **National Restaurant Association** and **Black Box Intelligence** peer benchmarks; reset KPI targets for the coming year.

## 30 / 60 / 90 Day Implementation

```mermaid
flowchart LR
    A[Day 0: Audit] --> B[Days 1-30]
    B --> C[Days 31-60]
    C --> D[Days 61-90]
    B -->|Install weekly P&L<br/>POS clean-up<br/>Inventory bookend| B1[Food + Labor visible weekly]
    C -->|Add RevPASH<br/>3PD audit<br/>Loyalty launch| C1[Throughput + retention visible]
    D -->|Menu engineering<br/>Scheduling model<br/>Quarterly scorecard| D1[Prime cost <65% sustained]
```

**Days 1-30 — Visibility.** Install **weekly P&L** with true inventory bookend. Clean up POS categories so food, beverage, and modifiers map correctly. Pull **30 days of DoorDash, Uber Eats, Grubhub** statements to compute real take rate.

**Days 31-60 — Throughput.** Build a **RevPASH dashboard** by day-part. Audit 3PD: drop the lowest-margin platform tier or renegotiate. Launch a **POS-tied loyalty program** (**Toast Rewards** or **Square Loyalty**) and capture phone/email at every check.

**Days 61-90 — Discipline.** Run menu engineering — **stars, plowhorses, puzzles, dogs**. Build a **forecast-driven scheduling model** (sales-to-labor ratio by half-hour). Publish the first **quarterly KPI scorecard** to the team. Lock prime cost below **65%** as the non-negotiable floor.

## FAQ

**Q: My food cost is "fine" at 35%. Why does my accountant say I'm losing money?**
Food cost at **35%** is high-end for full-service casual. Combined with full-service labor running **35-37%**, prime cost is **70%+** — which mathematically leaves nothing for rent, utilities, insurance, or profit. The accountant is right.

**Q: I do $1.8M in sales and feel like I'm working for free. What KPI explains it?**
Almost always prime cost above **65%**. On **$1.8M** revenue, every point of prime cost is **$18,000** of operator take-home. Three excess points = **$54,000** gone.

**Q: Should I drop DoorDash entirely?**
Only if **3PD is below 5% of your sales** and your dine-in is at capacity. Most independents are better served capping 3PD at **15-20%**, renegotiating to the **15% lite tier**, and pushing customers to first-party online ordering (**Toast, Square, Olo**).

**Q: How do I measure repeat-visit rate without a loyalty program?**
You can't, accurately. The minimum-viable substitute is **credit-card hashing** in modern POS systems (**Toast, Square, Lightspeed**) which can identify card-level repeats without an enrolled loyalty account. Real number still requires opt-in identity capture.

**Q: What's the single most-leveraged KPI to fix first?**
**Prime cost.** Every single dollar of prime cost reduction drops to the bottom line. A **2-point** improvement on a **$2M** restaurant is **$40,000** of profit that did not exist last quarter.

## Sources

- National Restaurant Association — **2026 State of the Restaurant Industry** report (restaurant.org/research-and-media/research/research-reports/state-of-the-industry).
- Restaurant Business Online / Datassential — **"The independent restaurant sector shrunk by 2.3% in 2025"** (nrn.com).
- Toast — **Restaurant Payroll Percentage Guide**, 2026 (pos.toasttab.com/blog/on-the-line/restaurant-payroll-percentage).
- Restaurant365 — **How to Calculate Prime Cost in a Restaurant** (restaurant365.com/blog/how-to-calculate-prime-cost-in-a-restaurant).
- 7shifts — **Restaurant Prime Cost Guide** and **RevPASH for Restaurants: Definition, Formula, and Best Practices** (7shifts.com/blog).
- Cornell University School of Hotel Administration — **RevPASH** research; Boston University RevPASH portal (bu.edu/revpash).
- Baker Tilly — **Prime Cost Target Tips** advisory note (bakertilly.com/insights/prime-cost-target-tips).
- TouchBistro — **10 Essential Restaurant Benchmarks** and **21 Restaurant Metrics** (touchbistro.com/blog).
- Independent Restaurant Coalition — **Third-Party Delivery Apps Federal Regulation Brief** (independentrestaurantcoalition.com/delivery_apps_2025).
- Rezku — **Third-Party Delivery Fees in 2026: What DoorDash, Uber Eats & Grubhub Really Cost Restaurants** (rezku.com).
- Restroworks — **Restaurant Customer Retention Statistics** (restroworks.com/blog/customer-retention-statistics-restaurants).
- Sweetgreen (NYSE: SG), Chipotle (NYSE: CMG), Texas Roadhouse (NASDAQ: TXRH), Cheesecake Factory (NASDAQ: CAKE), Wingstop (NASDAQ: WING) — **2024 Form 10-K** filings (sec.gov/edgar).

Was this helpful?

Related in the library