The year that did not line up.
Rebuilding every period, week and year-on-year comparison for a retailer moving onto a four-four-five trading calendar, so the reporting agrees with finance and the comparisons are genuinely like for like.
Speak with NicConsumer Retail Group
Client name and figures withheld for confidentiality.
The problem on the table.
A retailer's trading year is not the calendar year. Many move to a four-four-five pattern, where each quarter is two four-week periods and one five-week period, the year is fifty-two whole weeks, and every week starts on the same day. It aligns trading, finance and budgeting on one set of boundaries, which is why finance teams ask for it.
It also breaks every comparison built on calendar dates. A period no longer maps to a month, so a month-on-month view compares two different things and quietly calls it a trend. Year on year is worse: the same period a year earlier does not start on the same date, so a calendar comparison silently shifts the window by several days.
The client had restated prior years onto the new pattern, so the reporting had to be rebuilt to match it rather than approximated around it.
What forensic looked like here.
The trading week was made the atomic unit, because it is the only unit that is the same length in both years. A year-on-year comparison became the same weeks a whole number of weeks earlier, which lands on the same trading days in every year regardless of how the periods are grouped.
That mechanism was chosen over the simpler option of comparing period numbers, and the reason is worth stating. Where two years share the same pattern the two approaches agree. Where an earlier year carries a different phasing, comparing period numbers puts unlike windows side by side while the week shift stays correct, so the stronger method was kept even where the simpler one would have worked.
The calendar is generated from the client's own authoritative workbook by a committed script and never typed by hand, and the script validates its own output before it is accepted: the right number of days, the right number of weeks, every week starting on the trading week's own day, and no week straddling a boundary.
Every figure names both dated windows it compares, so a reader can see exactly which trading days sit on each side of it.
- A day-level calendar generated from the client's own source workbook
- Validation on every rebuild: week count, day count, week start, no split weeks
- Year-on-year on a whole-week shift, correct across differently phased years
- One calendar resolved by every surface that computes a period
- Both compared windows named on every figure
What changed.
Reporting periods, budget periods and trading weeks now share one set of boundaries, so the figures on the dashboard and the figures in the finance pack describe the same days.
The comparison basis is stated rather than assumed, and the calendar is regenerated from the client's own workbook rather than maintained by hand, so a future restatement is a script run instead of a reconciliation exercise.
- 4-4-5restated trading calendar
- One sourcethe client's own workbook
- Whole weeksthe year-on-year mechanism
- Validatedon every rebuild
Other engagements.
Your number is next.
A proper conversation with the person who would do the work. Tell me where you are and what you are weighing up, and you will get an honest, forensic read on what I would change, what I would leave alone, and where the real upside is.
Speak with NicThe senior read,
once a week.
Most marketing newsletters are noise on a schedule. This is not that. Once a week I send a short, considered read on the shifts that actually change what you should do, across the AI shift, PPC, technical SEO and analytics. Filtered, weighted and signed off by me before it goes out. The judgement is mine, and now you can hear every edition in my own voice.
Nic CraneFounder, The Crane Consultancy
Join readers of The Crane Standard.