U.S. retail and food-services sales retreated in July after advancing in June. The Census Bureau's new advance estimate places seasonally adjusted sales at $763.6 billion, down 0.6% from the previous month but 5.0% above July 2025. The agency's 90% confidence interval for the monthly change is plus or minus 0.4 percentage point, so the decline is statistically distinguishable from zero on that measure.
The headline is a current-dollar sales measure, not a count of cartons, store replenishments, parcels, truckloads or stops. Census adjusts for seasonal, holiday and trading-day patterns but not for price changes. FreightNews therefore treats the release as a demand signal rather than a direct measure of freight volume.
The category split was uneven. Nonstore retailers fell 2.2% from June and motor-vehicle and parts dealers fell 1.8%. Gasoline-station sales decreased 0.9%, while clothing and accessories stores rose 1.9%, health and personal-care stores rose 0.7%, food services rose 0.5%, and general merchandise stores rose 0.3%. Sporting-goods, hobby, musical-instrument and book stores were essentially flat on the adjusted estimate.
Removing motor vehicles and gasoline narrows the monthly decline to 0.2%. That comparison matters for freight desks because vehicle sales and fuel prices can move the total without producing the same pattern in dry-van, parcel, refrigerated or store-delivery demand. It still does not isolate physical units or transportation requirements inside the remaining categories.
The year-over-year picture is firmer than the one-month move. Total sales were 5.0% above July 2025, nonstore retail was 7.7% higher, and the May-through-July total was 6.3% above the same three months last year. Those nominal comparisons leave both price and product-mix effects in the data, so they do not cancel the caution around load counts.
The operating inference is that fall planning should not be built from the total alone. A monthly retreat led partly by autos and nonstore retail may soften some inbound, fulfillment and final-mile schedules, but July's gains in clothing, health, food service and general merchandise can support different equipment and delivery patterns. The release does not prove that peak season is either secured or lost.
Carriers and brokers should compare July with customer purchase orders, booked inbound containers, accepted tenders, order lines, cube per shipment, stop density and appointment calendars. Segment the review by retail category and lane. A national nonstore decline is much less useful than knowing whether a specific fulfillment center reduced outbound orders or merely changed the value and mix of what it shipped.
Capacity decisions also need timing discipline. The advance survey draws on an early sample of about 4,800 retail and food-service firms, and Census can revise the estimates as more complete data arrive. The agency has also announced a separate historical retail revision process later this year. Record the release vintage alongside any bid or staffing decision so a revised series does not overwrite the evidence available at the time.
For the freight desk, July calls for a narrower forecast, not a broader slogan. Hold equipment, labor and purchased transportation against live customer orders and lane economics, then use the federal category split to test where the book of business is diverging from the national signal.