NYC Clean Eats

The year-long A

The letter grade system gives a restaurant a second chance before a bad grade goes up on the window. We measured how long that second chance actually takes.

How the system is designed to work

When a restaurant is inspected, violations carry points, and a lower score is better. A score under 14 points earns an A, which goes straight into the window.

Score 14 or more points and something different happens: no new grade is issued. The restaurant is scored but ungraded while it waits for a re-inspection, and that second score becomes the official letter. The logic makes sense on paper. An inspection is a snapshot of one day, an inspector might catch a kitchen on a bad morning, and a business should get an opportunity to fix issues before a B or C hurts its trade.

What the rules do not say out loud is what happens to the placard while the restaurant waits. Because no new grade has been issued yet, the old A stays in the window.

What actually happens on a failed inspection

Here is that sequence with the real intervals from the data rather than a hypothetical.

DayWhat happensWhat is in the window
0 Inspected. Scores 24 points, including evidence of rats. Not enough to close. The A from the last cycle
152 Re-inspected. The median wait. Still failing, as 37% of follow-ups are. Still that same A
175 A B or C is finally issued. Median across restaurants that reach one. Grade Pending, or the new letter

Five months to the second look. Closer to six before a lower letter exists. At the top quartile it is 303 days, and for one restaurant in six it takes more than a year.

This delay is not a matter of restaurants breaking rules, because every step here follows official procedure. The issue is that a restaurant could fix its violations on any day in that period, and if it does not, nothing on the window placard informs the public.

The Grade Pending card and tribunal appeals

Even after an inspector completes a re-inspection and assigns a B or C score, the letter still might not appear in the window. City rules allow the owner to post a Grade Pending card instead while they challenge the citations at an administrative hearing through the Office of Administrative Trials and Hearings (OATH).

That appeal opens another window of its own. Owners present evidence, photos or testimony to dispute an inspector's observations or the point calculation, and the goal is to get enough points dismissed to bring the total back under 14 for an A. Scheduling a hearing and waiting for a written decision stretches the process out by weeks or months, and throughout it the public sees a generic card rather than the score or the violations found in the kitchen.

Between the wait for a re-inspection and the length of an appeal, a restaurant can operate for a substantial part of a year before a B or C ever hangs on its door.

Why clean A ratings get stale

There is a second effect, and it applies to kitchens that are doing nothing wrong. An A is posted on the day it is earned and stays there until the next inspection, so the letter ages quietly in the window.

We measured that too. Across 18,908 restaurants currently showing an A, comparing the date the city issued that grade against the date of the most recent inspection in the grading cycle:

Age of the A in the windowShare of restaurants
Median age288 days, about nine and a half months
Older than one year37.1%
Older than eighteen months17.2%
Older than two years6.2%

So more than a third of the A grades on display in New York were earned over a year ago. The letter is accurate about the day it was issued. It is silent about the twelve months since.

Those old grades are not evenly spread, though. Split them by whether the restaurant has been inspected again since:

Since that A was issuedRestaurantsMedian age of the AOver a year old
No further inspection 16,665 (88%) 260 days 29.4%
Inspected, no new grade issued 2,243 (12%) 629 days 95.0%

A restaurant that simply has not been visited again carries an A about eight and a half months old, which is roughly the inspection cycle working as intended. The other group is a different thing entirely. Their A is a median of 629 days old and 95% of them are over a year, even though an inspector was there a median of 87 days ago.

That is the first half of this article showing up in the second. A failing inspection issues no new grade, so the old A stays up and keeps ageing while the restaurant waits, and the longer the wait the older the letter in the window gets. Roughly one A placard in eight in New York has a more recent, ungraded inspection sitting behind it.

Stated as "no new grade has appeared" rather than "it failed", because that is what the record shows: a grade still being processed would look the same. A failing inspection issuing no grade is the designed behaviour rather than an exception, though. Counts only grading-cycle inspections, so an administrative or calorie-posting visit does not stand in for a look at the kitchen.

What the rules specify against what happens

The city is not vague about how the process should run.

Thirty days. That is the shape of the intended process: inspect, give the operator a month to put it right, come back and settle the grade.

0 500 1,000 1,500 30-day target median, 152 days 0-3061-90121-150181-210241-270301-330 days from the failing inspection to the next one
n = 12,822 failing initial inspections. A further 10.5% waited longer than 360 days and 3.3% were never followed up, neither of which is drawn.

The first bar is the tallest, and it is worth being careful about that. The single most common outcome is a follow-up inside the target month, and 12.5% of cases land there. It would be easy to look at that bar and conclude the system is working.

The median line is the correction. The distribution does not fall away after the first month, it flattens and keeps going: 11.3% in the second month, 9.2% in the third, still 5.6% in the eleventh. Half of everything sits to the right of 152 days. A further 10.5% waited more than a year, and those are not even on the chart.

Most common is not the same as typical. The target is met more often than any other single outcome, and it is still missed seven times out of eight.

Worth reading the audit itself, because the city already knew. Back in 2015 the Comptroller found that DOHMH "did not consistently attempt follow-up inspections in a timely manner, and exceeded the agency's informal 30-day target more than a quarter of the time". For establishments scoring 14 or more points, it found the mandated timing for beginning the next inspection cycle was exceeded approximately 50 percent of the time.

That audit is from 2015, so it does not describe today. What we can measure is today, using the city's own published inspection records. The intended month has become five months.

Summary of inspection wait times

Measured over 12,822 failing initial inspections, all dated on or before April 2025 so that at least sixteen months have passed and nearly every follow-up that will happen has.

MeasureValue
Median wait for the next inspection152 days
Followed up within the 30-day target12.7%
Took longer than six months40.4%
Never followed up at all3.3%
Follow-ups that also failed36.8%
Reached a posted B or C eventually36.1%
Median days to that B or C175 days
Upper quartile wait time303 days
Took over a year to resolve16.8%

Why this matters when checking a grade

An A placard shows that a kitchen met city standards on one specific date. It does not show when that inspection happened, or what inspectors found on any visit since. A restaurant with an A in the window today may have scored badly months ago and still be waiting for its follow-up, or be contesting the result at a hearing.

So read the date, not just the letter. Checking the inspection date and the points behind it tells you whether the grade describes the kitchen as it is now or as it was a year ago, and that is the difference the placard cannot show you. Our score does both: it shows what is behind the letter, how long it has been up, and whether the wait is already past the 152-day median.

Sources

Look up a restaurant · How the score is built · More insights