USDA Report 3228 is the weekly grocery store beef feature activity report, published by AMS every Friday afternoon. It counts the number of US grocery stores running an advertised feature on each beef cut; pork and chicken get the same treatment in companion reports (2868 for pork, 2756 for chicken). It is the most direct public read on where retail demand is being pointed, and one of the least-read datasets in the protein complex.
A wholesale buyer who reads 3228 weekly sees where retail demand is being steered before it shows up anywhere else in public data. The report is the earliest public record of retail feature commitments.
What the report actually measures
3228 surveys upwards of 29,000 US grocery stores each week, with the exact base varying by chain participation. For each protein category and each specific cut, it counts:
- Total number of stores featuring the cut - Average advertised retail price per pound - Week-over-week change in store count - Year-over-year change in store count
A "feature" is an item appearing in a store's weekly advertised circular at a promotional price. Featured cuts move multiples of the volume of unadvertised cuts at the same store, which makes the store count a direct measurement of how much retail-channel demand is being pulled into the next week's consumer purchases.
The report breaks out each protein by cut group and then by individual item. On the beef side that means rib, loin, sirloin, chuck, round, brisket, ground beef, and deli groups, with the specific cuts (ribeye, strip, tenderloin, ground beef by lean point, chuck roast, top round, and more) listed inside each. Each line is a tradable signal.
Why the timing matters
A retailer commits to a feature program weeks ahead of the in-store week, and the wholesale buy behind it happens before the ad runs. So the week a cut's store count jumps in 3228 is not the week the wholesale print gets bid. It is the week the buying is done.
The feature history since 2018 shows it. Take the weeks when a cut's conventional store count jumped well above its own recent run, then follow the cut's wholesale print measured against its cutout, so the season drops out:
| Feature surge in | Surge weeks | Print two weeks later, vs the cutout | Weeks that softened |
|---|---|---|---|
| Tenderloin | 33 | -2.0% | 73% |
| Boneless top sirloin steak | 19 | -3.4% | 74% |
| Bone-in ribeye roast | 37 | -2.6%, and -7.7% four weeks on | 51% |
| Spareribs | 24 | -6.4% | 75% |
| Bone-in loin chops | 21 | -2.5% | 67% |
| Sliced bacon | 7 | -5.4% | 86% |
The steak cuts, ribs and bellies give back after the surge week, because the chains already own the product. Ground beef is the one with a visible run-up: the 81s print rose in the three weeks into a top ground beef feature week in 11 of 13 cases, then gave back over the next two in 9 of them. Pork butts run up into a feature week too, against the cutout in 17 of 25 cases, and hold. Round roasts, boneless ribeye steak and chuck roast show no consistent pattern on either side of the surge.
So the read on a big feature week is not "the wholesale price goes up next". The buy is in the print. The question is how hard the cut fades once the window ends, and whether supply is tight enough to hold it.
What the signals look like
A few patterns are particularly actionable.
Ribeye runs near 4,500 featuring stores in a typical week of the 2023 to 2026 survey history. A print past roughly 10,000 stores puts the week in the top fifth of all weeks, and the big holiday pushes (Father's Day, July 4, the December window) have printed past 20,000. A top-quintile ribeye week says retailers are leaning into a middle-meat pull. The wholesale ribeye bought for those programs is already in the print, and the bone-in ribeye has given back against the cutout in the weeks after a surge more often than not.
Ground beef runs near 20,000 featuring stores in a normal week. Past roughly 26,000 the week is in the top fifth, and the biggest grind pushes on record have printed into the 40,000s and above. A top-quintile ground beef week is a broad grind program whose buying has run the 81s up into that week; the print gives back after it, and the trim ladder (81CL, 65CL, 50CL) carries the same shape. Ground beef features cluster around Memorial Day, Labor Day, and the back-to-school weeks of late August.
Pork loin features run in a less predictable cycle than beef because pork retail promotion is more sensitive to packer-side promotional dollars, which fluctuate quarter to quarter. The usable read is the same shape: a week that jumps well above the cut's own recent range is a whole-loin push whose buying is done. Bone-in chop surges were followed by a softer bone-in loin print two weeks on in two of three cases; boneless loin surges showed no consistent direction.
Chicken breast is the biggest single line in the report: a normal week runs near 33,000 featuring stores, and a surge past roughly 40,000 puts the week in the top fifth. During football season, wings take over the chicken feature category and breast counts compress; the wing count is the relevant signal in that window.
What week-over-week deltas tell you
The absolute store count matters, but the week-over-week delta is often a stronger signal. A cut that roughly doubles its featuring stores in a single week (say a ribeye print jumping from 4,500 to 9,000) is showing accelerated retailer commitment, which usually means the feature program is responding to a specific cost or supply event. A cut that gives back half its count week over week is showing retailer pullback; with no ad behind it, the cut has to find its bid elsewhere.
The classic setup is a category showing simultaneous week-over-week store count surge AND year-over-year store count gain. Both directions confirming means the retail pull is genuine and not just a calendar effect (a feature week shifting between calendar weeks year over year).
Where 3228 misleads
The report is honest but it has two built-in limitations a buyer needs to factor in.
First, it surveys a self-selecting sample of US grocery chains. Some major chains participate consistently, some participate sporadically, and a handful never report. Walmart's participation pattern in particular has shifted over the years. The headline store counts therefore measure feature activity at participating chains, not US grocery total. A change in participation can swing the weekly number in ways that don't reflect actual retail behavior. The week-over-week trend within a stable participation set is more reliable than the absolute level across years.
Second, the report measures feature commitment, not feature execution. A store that commits to a ribeye feature can run out of inventory mid-week and reduce the actual sell-through below what the feature count implies. In tight-supply conditions this happens more often, which is one reason the price response to a feature week varies so much by cut.
When the report is most useful
The ads land in the holiday week itself. Since 2018 the ham store count has peaked in Easter week every year, and the ribeye count in Christmas week or the one before. The wholesale print tops out earlier: against its cutout, the tenderloin has peaked a median three weeks before its Christmas feature peak, the ham four weeks before Easter, spareribs about six weeks before July 4. By the time the ad page is full, the wholesale top is usually behind you.
3228 is most actionable during three specific windows in the calendar:
April through May for the Memorial Day cycle. Watch ribeye, strip, brisket, and ground beef counts from mid-April; the counts tell you which cuts the chains committed to, and the Memorial Day grind features are the ones the 81s run up into.
Mid-June through early July for Father's Day and July 4. Same cuts, compressed timeline.
Mid-August for Labor Day. Slightly different cut mix because Labor Day grilling is more focused on grind, hot dogs, and entry-level cuts than the higher-dollar middle meats of Memorial Day and July 4.
Outside these windows the report is still informative but the moves are smaller and the signal-to-noise is lower.
Reading 3228 on Meat Read
Meat Read's Retail Features surface tracks the full feature-report family weekly (3228 for beef, with the pork and chicken companions), with week-over-week and year-over-year deltas highlighted by cut. The activity index aggregates the report into a single readable number per protein so the week's overall promotional pulse is visible at a glance, and the per-cut detail lets a buyer drill into the specific signals that affect their book. Cross-referenced against the cut's own price page, the gap between the feature peak and the wholesale top is visible in real time.