How Restaurant Health Grades Are Assigned and What They Miss
A letter posted in a restaurant window feels like a verdict. A means safe, B means acceptable with caveats, C means think twice.
In reality, that single letter compresses a multi-hour inspection, dozens of weighted violation codes, and a regulatory philosophy that varies from city to city into a symbol most diners read in under two seconds. The grade is useful.
Trending Now!!:
It is also, by design, incomplete: it measures conditions observed during one visit that typically lasts one to three hours, out of thousands of hours a kitchen operates each year, and it says almost nothing about what happens on the other 364 days.
Grades are assigned by trained inspectors who tour a kitchen, dining room, and storage areas, tally point deductions against a list of code violations weighted by public health risk, and convert the resulting score into a letter, number, or pass/fail designation, depending on the jurisdiction.
What the system misses is just as consequential as what it catches: employee illness that never surfaces during a scheduled walkthrough, inspection frequency that thins out for restaurants with strong track records, and a scoring model that treats a single kitchen visit as a stand-in for year-round practice.
The Regulatory Backbone: One Federal Code, Dozens of Local Interpretations
Nearly every state and local grading program in the United States traces its violation categories back to a single source: the FDA Food Code, a model set of guidelines that identifies 56 different inspection items and directs health departments on what to look for and how heavily to weigh each finding.
About half of those 56 items qualify as critical violations, meaning they meaningfully raise the risk of foodborne illness and carry disproportionate weight when a final score is calculated.
That shared foundation is where the uniformity ends. States adopt the Food Code on their own timelines, adapt it to local conditions, and then layer wildly different presentation formats on top of it. Some states use traditional letter grades of A, B, or C, while others rely on numerical scores or colour-coded systems.
Florida posts percentage scores alongside letter grades, and many Texas counties operate on a straightforward pass or fail basis. Massachusetts deducts points based on both the severity and frequency of violations, while Nevada weights its system heavily toward penalizing critical infractions.
New York City runs the most closely watched version of the letter-grade model. A restaurant that scores 0 to 13 points during its inspection earns an A, a score of 14 to 27 points earns a B, and 28 points or more results in a C, with grade cards required to be posted where passersby can easily see them.
The city’s process also builds in a second chance that many diners never notice: if a restaurant scores 14 or more points on its initial unannounced inspection, it is not graded immediately. A follow-up inspection is scheduled within four to eight weeks, and it is that second visit, not the first, that determines the letter ultimately posted. A restaurant displaying Grade Pending is not hiding a failing score. It is mid-appeal, contesting violations at an administrative hearing before a final grade goes up.
Los Angeles County popularized letter grading on the West Coast, and its point logic runs in the same direction as New York’s: fewer points mean a cleaner inspection. Arizona’s Maricopa County, by contrast, treats grading as opt-in.
The health department asks the person in charge at the time of inspection whether the establishment wants to participate in the grading system, and an operator who declines gets a Not Participating designation on the public report instead of a letter. A diner scanning county records for a Maricopa restaurant and finding no grade at all is not necessarily looking at a clean operation. They may be looking at one that opted out.
The upshot for anyone comparing restaurants across city lines: a B in Fishers, Indiana carries different math than a B in Manhattan, and a percentage score in Florida cannot be dropped into a Massachusetts point-deduction framework without translation. National grading comparisons that treat these systems as interchangeable are, functionally, comparing different instruments and reporting a single unit.
What Inspectors Actually Look For, and How Points Accumulate
Inspectors work from a checklist that spans food temperature control, cross-contamination risk, handwashing facilities and practices, pest evidence, food source verification, and employee health reporting. Violations get sorted into two tiers. Critical violations are the ones with a direct line to foodborne illness: improper hot or cold holding temperatures, evidence of pests, contaminated food, inadequate handwashing setups.
In New York City’s model, critical violations run five to nine points each, while general violations covering administrative and structural issues, such as inadequate lighting or missing required postings, run two to five points each. The most frequently cited critical violations in the city involve food held at unsafe temperatures, signs of mice or cockroaches, and handwashing failures.
The industry misconception worth correcting here: operators and diners alike often assume a grade reflects overall cleanliness. It does not, at least not primarily. A restaurant can have spotless floors and a beautifully organized walk-in and still take a C if a single reach-in cooler is running six degrees too warm during the walkthrough.
Conversely, a kitchen with cosmetic clutter but airtight temperature logs and proper cooling procedures can post an A. The grade is a risk score, not a tidiness score, and conflating the two is one of the most common reading errors among consumers.
Some violations that carry outsized real-world risk are also the hardest to catch on a scheduled visit. An employee who shows up to work sick is one of the most common vectors for spreading foodborne illness, but inspectors largely depend on operators to foster a culture in which staff feel comfortable self-disclosing symptoms rather than working through them.
There is no reliable mechanism for an inspector to detect, during a single visit, whether a line cook came in with a stomach bug three days earlier and never told anyone.
What the Grade in the Window Cannot Tell You
This is where the gap between perception and reality widens, and where most consumer-facing coverage of restaurant grading stops short.
Inspections are a snapshot, not a stream:
Research examining San Francisco’s inspection program found that routine health inspections cover less than 1 percent of a restaurant’s annual operating time, with the city’s health department inspecting each establishment only two or three times per year at most.
That is not a San Francisco quirk. It is structurally true of nearly every jurisdiction, because most health departments determine how frequently to reinspect a restaurant based on its prior grade, meaning better-performing restaurants get inspected less often.
A restaurant that earns an A in March and is not due for reinspection until the following year could shift staffing, change suppliers, or lose a manager who enforced cooling procedures, all without that change ever registering on the public record.
Favorable scores can reduce future scrutiny:
In San Francisco’s risk-ranking model, a restaurant scoring above 80 out of 100 points is not scheduled for another inspection for the rest of the year. That policy makes operational sense from a resource-allocation standpoint, since departments have finite inspector staffing.
It also means the grade posted in the window can be, at any given moment, many months stale, and the public has no easy way to know how stale without checking inspection dates in a jurisdiction’s online database.
Complaint systems fill gaps, unevenly:
When a health department is not physically inside a restaurant, its main window into ongoing risk is consumer complaints. A national survey of local health agencies found that 91 percent maintain a system allowing the public to report suspected foodborne illness tied to a specific restaurant.
But complaint systems are voluntary, underused relative to actual illness incidence, and dependent on consumers correctly attributing symptoms to a specific meal, something epidemiologists know people do poorly given incubation periods that can run from hours to weeks.
Research linking online review language to inspection outcomes has tried to close that gap by mining platforms like Yelp for illness-related complaints, but even that approach runs into selection bias toward more active reviewers and timing mismatches, since a review describing illness may appear weeks or months before or after the inspection it is compared against.
The stakes are rising, not falling:
A 2025 report tracking food safety trends found confirmed foodborne illness cases climbed from 1,118 in 2023 to 1,392 in 2024, while hospitalizations from confirmed outbreaks more than doubled, jumping from 230 to 487 over the same period.
That trajectory matters for how much weight the public should place on grading systems as a protective measure. Grading correlates with improved outcomes, but it does not eliminate the underlying risk curve, and treating a posted A as a guarantee misreads what the letter is actually certifying.
Grading works, within its limits:
It would be a mistake to read the above as an argument that letter grades are theater. New York City, which marked fifteen years of restaurant grading in July 2025, saw measurable improvement almost immediately after the program launched in 2010: fewer restaurants cited for unsafe holding temperatures, and higher enrollment in food worker safety training.
Broader survey research has found that health agencies which grade restaurants and routinely disclose the results to the public report fewer foodborne illness outbreaks than agencies that skip grading or keep results private. The system changes operator behaviour because it is public and because it carries reputational and financial consequences. It just does not, and structurally cannot, function as continuous surveillance.
Reading a Grade Like Someone Who Understands the System
For a diner deciding where to eat, the letter grade is a reasonable first filter but a poor final answer. A more useful habit is checking the underlying inspection report rather than the grade alone, since most jurisdictions now publish itemized violation histories online.
A restaurant with an A and a clean multi-year history is a meaningfully different bet than a restaurant with an A earned on a second-chance reinspection after an initial score in the C range. Reinspection narratives that describe a violation as corrected reflect a different operational reality than ones flagged as a repeat offense, and that distinction never shows up in the letter itself.
For operators, the practical lesson is similar to what separates consistent A performers from restaurants that scramble before every visit. The restaurants that hold their grades over time tend not to be doing anything exotic.
They build food safety into daily operations rather than preparing for it right before an inspector arrives, often working from a formal hazard analysis framework that identifies where risk exists in the operation and documents exactly how it gets controlled.
Because reinspection frequency is often tied to prior performance, a single strong inspection can buy a restaurant months of reduced scrutiny, which raises the stakes on whether that snapshot genuinely reflects ongoing practice or simply good timing.
The letter in the window will keep functioning as shorthand, and it should. It condenses genuine regulatory rigour into something a passerby can read in an instant.
But the honest way to use it is as a starting point rather than a conclusion: a signal that a kitchen passed a specific test on a specific day, inside a system that, by its own design, only ever sees a fraction of what actually happens behind the line.
