How to phase a season plan by month (without copying last year)

The season total gets all the attention. The monthly phasing gets a copy-paste. A planner sets the annual number with care, argues it in the review, defends the margin rate, and then splits it into months by dropping last year's monthly percentages onto this year's total. The phasing, which decides when the money is supposed to arrive, is treated as a formatting step.

It is not a formatting step. Phasing is a forecast of its own, and copying last year's shape is a forecast that assumes nothing changed: same calendar, same weather, same promotional cadence, same channel mix, same launch timing. None of that is true, so the phased plan is wrong in a specific and expensive way. It tells you to expect revenue in months it will not arrive, and to be surprised by revenue in months you did not plan for.

Nobody calls this a mistake because it never shows up as one. A phasing error does not stock you out and it does not leave you with obvious overstock. It leaks quietly: you promote in a soft month because the plan said the month should be strong, you under-buy for a peak the plan flattened, and at year end the total looks close enough that nobody investigates the shape underneath it.

Start with what last year's shape actually encodes. The monthly percentages you are pasting are the record of a specific year: the year Easter fell in March, the year the heatwave pulled summer forward three weeks, the year you ran an unplanned promo in October to clear a supplier overhang. All of that noise is baked into the percentages, and the copy-paste treats it as signal.

This year the calendar has moved. Easter is in a different month. Peak trading days land on different weekends. A launch that shipped in February last year slips to March this year, so the demand it drives moves a full month, but the phased plan still expects it in February. Every one of those shifts is a phasing error, and they do not cancel out. They stack up into a monthly shape that is confidently wrong.

The tell is the monthly variance report. Every month is off plan, some high and some low, and every month there is a story: the weather, the promo, the launch, the comp. When every month needs a story to explain the miss, the plan was not phased to demand. It was phased to a memory of last year, and last year is not coming back.

Worth naming the two ways the copy-paste goes wrong, because they are different problems. The first is calendar shift: fixed dates like Easter, Black Friday and the return-to-work weeks land in different months year to year, so a percentage tied to March last year is tied to the wrong demand this year. That one is at least predictable, and it is the one brands sometimes correct. The second is structural drift, and it is the one they miss: the channel mix moved, the launch cadence changed, a category you barely had last year is now a third of the range. Last year's percentages encode last year's structure, and structure does not repeat even when the calendar does. Correcting for the calendar and ignoring the structure gets you a plan that is confidently wrong in a more sophisticated way.

Accuracy in the phasing is what stops the monthly firefighting. When the shape of the plan matches the shape of demand, the variance report gets short and boring, which is exactly what a variance report should be. The months stop needing stories because the months stop being surprises.

Trace where a phasing error lands and it always ends at margin. Phase a peak into the wrong month and the buy follows the plan: you receive the stock early, it sits waiting for demand that has not arrived, and by the time the real peak lands you are either stocked out because you already discounted the early sitting stock, or you are carrying it into the peak at a markdown you took to move it.

Phase demand out of a month it will actually appear in and the mirror problem happens. The plan says the month is soft, so you thin the buy, so you stock out in the middle of unexpected demand and lose the full-price sales you could most profitably have made. Either direction, the phasing error becomes a margin event, and it was scheduled the day you pasted last year's percentages.

The reason it stays invisible is that phasing errors hide inside an accurate total. You can hit the annual number to the dollar and still have lost margin in every month along the way, over-promoting the soft months and under-buying the strong ones. The total is a report card that grades the wrong thing. The shape is where the money is made or lost.

There is a second-order cost that makes it worse. A phasing error does not just misprice one month, it corrupts the read on the next one. If the plan expected a peak in April and the peak actually lands in May, April looks like a miss and May looks like a beat, and the planner reacts to both: they promote into the soft April to hit the number, pulling forward demand that would have arrived anyway, and then they are surprised by a May that the plan said should be quiet. The reaction to the first phasing error creates the conditions for the next one. A plan phased to demand does not generate these false signals, so the planner stops chasing ghosts month to month.

Ten points of inventory sitting past its selling window is cash you scheduled into markdown at planning time, months before a single unit shipped. It is also the cash you need for the next buy, which is how a bad phase in one season quietly finances caution in the next.

The alternative is to phase the plan the same way you should forecast anything: from demand. Build the monthly shape from the live demand model, aligned to this year's trading calendar, adjusted for where the launches actually land and how the channels are mixing now, with the promotional plan as an input rather than a ghost baked into last year's percentages. The season total stays owned by the planner. The shape underneath it is computed, not copied.

Then keep it live. A phasing that is right at sign-off drifts the same way any plan drifts, so the model re-phases against real sell-through through the season. When a peak comes early, the plan moves the demand forward and the buy follows before the reorder window closes. When a month softens, the plan pulls the expectation down and finance sees the working-capital impact the same hour, not at the quarter-end reconcile. Every re-phase is a snapshot in the audit trail, so the shape of the plan and the reason it changed are always on the record.

Copying last year's monthly shape is not phasing. It is scheduling this year's markdowns before the season has started.

Phase from demand and the monthly variance report stops being a monthly confession. The shape of the plan matches the shape of the business, the months stop needing stories, and the margin you used to leak into the wrong months stays where you planned it.