What the 4-5-4 calendar is
The 4-5-4 calendar is the retail industry's fiscal calendar, which divides each quarter into a 4-week month, a 5-week month and a 4-week month so that reporting periods always start and end on the same weekday. The National Retail Federation, which publishes it, describes it as a guide "that ensures sales comparability between years by dividing the year into months based on a 4 weeks - 5 weeks - 4 weeks format," and says the layout "lines up holidays and ensures the same number of Saturdays and Sundays in comparable months." Walmart's fiscal calendar aligns with the NRF model: its fiscal 2027 runs from 1 February 2026 to 31 January 2027.
The first time this bit me I was reconciling a KeHE file to a retailer's own period report and could not find $40,000. There was no missing money. The retailer's March had five weeks in it and mine had four.
How the months line up
The retail year starts in what the calendar calls February and ends in what it calls January. Twelve months of 4, 5 or 4 weeks each is 52 weeks, so the pattern repeats four times.
| Quarter | Month 1 | Month 2 | Month 3 |
|---|---|---|---|
| Q1 | February (4wk) | March (5wk) | April (4wk) |
| Q2 | May (4wk) | June (5wk) | July (4wk) |
| Q3 | August (4wk) | September (5wk) | October (4wk) |
| Q4 | November (4wk) | December (5wk) | January (4wk) |
So the answer to "how many weeks in April" is four, on this calendar, every year. March is five. That single fact is behind more bad month-over-month reads than any other feature of the system.
Take a nine-store grocery chain's center-store department.
| Retail month | Weeks | Sales | Sales per week |
|---|---|---|---|
| February | 4 | $4,100,000 | $1,025,000 |
| March | 5 | $5,200,000 | $1,040,000 |
| April | 4 | $4,200,000 | $1,050,000 |
March reads as up 27% on February ($5.2M against $4.1M). Per selling week it is up 1.5% ($1,040,000 against $1,025,000). April then reads as down 19% on March and is actually up 1.0% per week. Every real move in that department is inside a two-point band, and the headline swings 46 points across two comparisons. On a 4-5-4 calendar, divide by weeks before you react.
4-5-4 versus 4-4-5
The variants differ only in where the five-week month sits inside the quarter. 4-5-4 puts it in the middle. 4-4-5 puts it at the end, which pushes more volume into the quarter's final month and suits a business that closes hard at quarter end. 5-4-4 loads the front. The NRF maintains 4-5-4 as the standard for US retail, and it is what most large chains and their suppliers run, so a supplier reporting to several retailers will usually find one calendar covers most of them and a handful of accounts sit on their own.
The practical rule is not to guess. Ask each account which structure and which year-start it uses, then store the mapping once. A period label like "P3" is meaningless without it, and two accounts can both say "March" and mean date ranges a week apart.
The 53rd week, and the silent year shift
52 weeks of 7 days is 364 days, so the calendar drifts a day a year and more in a leap year. The NRF's rule: after laying out the 52-week calendar for a given year, if four or more days are left in January during the 53rd week, a 53rd week is added. NRF names 2006, 2012, 2017 and 2023 as 53-week years, which is roughly every five to six years. For comparability the NRF restates a 53-week year in the following year by pushing each week back one week and ignoring the first week of the fiscal year.
Two things break if you do not handle this.
The annual total is not comparable. A 53-week year carries an extra week of sales for free.
| Year | Weeks | Sales |
|---|---|---|
| FY2023 | 53 | $54,600,000 |
| FY2023 week 53 | 1 | $950,000 |
| FY2024 | 52 | $54,000,000 |
Reported, FY2024 is down 1.1% on FY2023. Strip the 53rd week and the comparable FY2023 base is $53,650,000, against which FY2024 is up 0.7%. Same two files, a 1.8-point swing, and the difference is one week in late January that only one of the years contained.
Every week after the shift is compared to the wrong week. Once the calendar absorbs an extra week, week 27 of this year sits against a week 27 of last year that covered different calendar dates. For most weeks that is noise. Around a holiday it is not: if Independence Day falls in week 22 one year and week 23 the next, the year-over-year read for both weeks is wrong, one up and one down, and neither is a demand signal. That is the same trap that shows up around tentpole events, and it is why NRF's realignment rule exists rather than being left to each retailer.
Where Scout fits
POS arrives at weekly grain, and weekly grain is what makes a retail calendar tractable: weeks are the atom, months and quarters are sums of them. Scout aligns the weekly POS it reads to the period structure you tell it you run, so a 5-week March is a 5-week March and a year-over-year comparison lands on the matching retail week rather than the matching Gregorian date. Scout reports against those periods. It is not your general ledger and it does not run your accounting close. If your fiscal structure and your GL disagree, the GL wins and Scout should be pointed at the same definition.
The short version
- The 4-5-4 calendar divides each retail quarter into 4-week, 5-week and 4-week months so like days compare to like days. April is always four weeks, March always five.
- Month-over-month comparisons on this calendar are meaningless until you divide by selling weeks. A 27% "gain" was 1.5%.
- 52 weeks is 364 days, so a 53rd week is added roughly every five to six years. NRF names 2006, 2012, 2017 and 2023.
- A 53-week year overstates the annual total and shifts every week's year-over-year partner. Restate before you report.
Sources: NRF, "4-5-4 Calendar"; 8th & Walton, "Walmart 4-5-4 Calendar Explained for Suppliers".