Year on year, minus eleven days

Year on year is the most used comparison in this region, and the least honest.

Every monthly pack carries it. This month against the same month last year, in green or in red, with an explanation underneath. In a market where the largest commercial month of the year moves about eleven days earlier each year, that comparison is mostly measuring the calendar.

Ramadan is expected to begin around the start of March. Because the month moves earlier every year, last year's began later in March than this one is expected to. So February will be set against a February that held a different amount of run-up, and March against a March that held fewer days of the month itself. Neither comparison is wrong exactly. It is that the largest variable inside both of them is nothing the team did.

What that hides runs in both directions. A February that looks strong may only be holding shopping that fell in March last time. A team that has a genuinely good March can be told it was flat, because last year's March was a different month wearing the same name.

The comparison worth having is aligned to the month, not the calendar. Day one against day one. The four weeks before against the four weeks before. Eid against Eid, allowing for the sighting. Everyone who works here knows this, and almost nobody builds it, because building it takes a morning and there is always something more urgent than a morning. I build mine in January, and not out of rigour.

January is when it is cheap. Nothing has happened yet, nobody has a result to defend, and a second column added now reads as method. The same column added in April reads as an excuse, because by April it is one.

Whoever builds it this month owns the story of the year.