Standard vs. Non-Standard Calendars

Introduction: Beyond the Gregorian Calendar

In many Financial Planning & Analysis (FP&A) functions, you’ll encounter a calendar_type argument. This might be confusing, as most people assume there is only one type of calendar!

To accommodate the unique operational and analytical needs of certain industries (particularly retail and manufacturing), ti often involves two primary calendar systems:

Standard Calendar (Gregorian)
Non-Standard Calendars (e.g., 5-4-4, 4-4-5, 4-5-4)

This guide explains the differences between these calendars and their typical use cases in business analysis.

1. The Standard (Gregorian) Calendar

This is the conventional calendar used for general financial reporting, scheduling, and external communication.

Structure & Use

Period Structure: 12 months of varying lengths (28–31 days).
Days/Weeks: 365 days per year (366 in a leap year).
Advantage: It matches external reporting deadlines, regulatory timelines, and standard corporate schedules.

Disadvantage: Inconsistent Comparisons

The primary drawback is that comparing operational metrics (like sales or margin) month-over-month (MoM) or quarter-over-quarter (QoQ) is challenging. The varying number of days or weekends leads to misleading conclusions.

For example, comparing sales from a 31-day month to a 28-day month will inherently show a “drop” in total revenue, even if the daily performance was identical. This makes true performance comparison difficult.

2. Non-Standard (Retail/Fiscal) Calendars

Non-standard calendars are explicitly designed to control for period-to-period variance. They achieve this by ensuring every comparable period has an identical number of weeks and, critically, an identical number of weekends. This allows for a cleaner “apples-to-apples” comparison of operational performance.

The most common types are the Retail Calendars (used in retail, restaurant, and manufacturing), such as 5-4-4, 4-4-5, and 4-5-4.

Example: The 5-4-4 Calendar Structure

The 5-4-4 structure ensures that every quarter is exactly 13 weeks long, and the month structure repeats perfectly:

Basis: Weeks in a period, not days in a month.
Total Year: 52 weeks (364 days).

Quarter Structure	Weeks	Days	Purpose
Period 1	5 Weeks	35 days	Captures seasonal spikes or month-end activity.
Period 2	4 Weeks	28 days	Standard monthly period.
Period 3	4 Weeks	28 days	Standard monthly period.
Total Quarter	13 Weeks	91 days	Ensures exact period alignment QoQ.

The 53rd Week Challenge

Since a 52-week year is 364 days, and the solar year is $\approx 365.25$ days, a 53rd week must be added every 5 to 6 years to keep the fiscal year aligned with the actual calendar year. This 53-week year is a crucial planning consideration for businesses using non-standard calendars.

Calendar Comparison

Feature	Standard (Gregorian) Calendar	Non-Standard (e.g., 5-4-4) Calendar
Basis of Period	Days in month	Weeks in period
Month/Period Length	Varies (28–31 days)	Fixed (4 or 5 weeks)
Consistency MoM/QoQ	Low: Days/weekends vary	High: Weeks/weekends are identical
Best For	General reporting, external financial statements (GAAP/IFRS)	Operational Analysis: Retail sales, inventory, labor costs, and year-over-year comparisons.

How to reference in `ti` functions?

The calendar_type argument is how you instruct the function on which system to use for its internal calculations and indexing.

To use a traditional calendar, pass: "standard"
To use a retail calendar, pass the respective structure: "544", "454", or "445"

Under the hood, ti handles the complexities: it generates the appropriate date keys and ensures that all your metrics are indexed correctly to the start and end of the chosen calendar’s periods.

Fiscal Year Start

By default, the retail calendar’s fiscal year starts nearest to January (fiscal_year_start = 1). Many retailers operate on a fiscal year that begins in February (e.g., NRF convention). You can control this with the fiscal_year_start parameter:

fiscal_year_start = 1 – fiscal year starts nearest to January 1st (default)
fiscal_year_start = 2 – fiscal year starts nearest to February 1st (common in US retail)
fiscal_year_start = 7 – fiscal year starts nearest to July 1st

The start date is determined by finding the week_start weekday (default: Sunday) nearest to the 1st of the specified month.

Examples

Year-to-date with a 4-4-5 retail calendar

The fiscal year starts nearest to February. Each quarter has 13 weeks split into 4-week, 4-week, and 5-week months.

contoso::sales |>
  ytd(
    .date = order_date,
    .value = margin,
    calendar_type = "445",
    fiscal_year_start = 2
  ) |>
  calculate() |>
  collect()

date	year	quarter	month	week	day	margin	ytd_margin	missing_date_indicator
2021-05-18	2021	2	4	16	18	406.840	406.840	0
2021-05-19	2021	2	4	16	19	711.351	1118.191	0
2021-05-20	2021	2	4	16	20	1424.101	2542.292	0
2021-05-21	2021	2	4	16	21	11338.631	13880.923	0
2021-05-22	2021	2	4	16	22	5358.767	19239.690	0
2021-05-23	2021	2	4	17	23	0.000	19239.690	1
2021-05-24	2021	2	4	17	24	0.000	19239.690	1
2021-05-25	2021	2	4	17	25	792.933	20032.623	0
2021-05-26	2021	2	4	17	26	74.550	20107.173	0
2021-05-27	2021	2	4	17	27	1432.710	21539.883	0

Notice that the year, quarter, month, and week columns now reflect the fiscal periods from the 4-4-5 calendar rather than the Gregorian calendar.

Month-to-date with a 5-4-4 retail calendar

contoso::sales |>
  mtd(
    .date = order_date,
    .value = margin,
    calendar_type = "544",
    fiscal_year_start = 2
  ) |>
  calculate() |>
  collect()

date	year	month	quarter	week	day	margin	mtd_margin	missing_date_indicator
2021-05-18	2021	4	2	16	18	406.840	406.840	0
2021-05-19	2021	4	2	16	19	711.351	1118.191	0
2021-05-20	2021	4	2	16	20	1424.101	2542.292	0
2021-05-21	2021	4	2	16	21	11338.631	13880.923	0
2021-05-22	2021	4	2	16	22	5358.767	19239.690	0
2021-05-23	2021	4	2	17	23	0.000	19239.690	1
2021-05-24	2021	4	2	17	24	0.000	19239.690	1
2021-05-25	2021	4	2	17	25	792.933	20032.623	0
2021-05-26	2021	4	2	17	26	74.550	20107.173	0
2021-05-27	2021	4	2	17	27	1432.710	21539.883	0

Grouped analysis with a retail calendar

Groups work exactly as they do with the standard calendar. Each group gets a complete fiscal calendar.

contoso::sales |>
  group_by(store_key) |>
  qtd(
    .date = order_date,
    .value = margin,
    calendar_type = "445",
    fiscal_year_start = 2
  ) |>
  calculate() |>
  collect()

Comparing calendar types

Below is a side-by-side comparison of how the same date is classified under a standard calendar vs. a 4-4-5 retail calendar:

Table 1

Feature	Standard (2024-03-15)	4-4-5 FY starts Feb (2024-03-15)
Year	2024	2024
Quarter	1	1
Month	3	2
Week	11	7

The retail calendar shifts the fiscal boundaries so that every comparable period contains the same number of weeks, enabling cleaner period-over-period comparisons.