How often should a merchandise plan be reforecast?

Monthly is the minimum for any business with a seasonal calendar, and weekly in season for direct-to-consumer channels where sales data arrives continuously and receipts can still be moved. The correct cadence is derived from the last point at which a decision can still change rather than from the reporting calendar — reforecasting more often than decisions can be made produces meetings, and less often produces reports about a season that is already settled.

What is the difference between a forecast and a reforecast?

A forecast is an estimate of what will happen. A reforecast replaces the estimated portion of a plan with actuals for periods that have closed and re-derives everything downstream of them. The important part is the re-derivation: updating actuals without re-solving the inventory targets and the receipt plan produces a document that is accurate about the past and unchanged about the future, which is the half that decisions depend on.

What has to be re-solved in a reforecast?

Everything downstream of the change. Replace actual sales, markdowns and receipts for closed periods, then re-solve the remaining periods in order: revised sales phasing, then inventory targets from the stock-to-sales ratio, then receipts from the inventory identity, then the open-to-buy position. Patching only the period that moved leaves the season total intact and every downstream target wrong, and because each patch leaves the periods after it out of line with the inventory identity, the errors stack until the plan stops reconciling to itself.

Should a reforecast change the season total or just the shape?

Both are legitimate outcomes and they lead to opposite decisions, which is why they have to be distinguished explicitly. A season running behind because demand is lower is a smaller season and the response is to cut receipts. A season running behind because deliveries landed late is a differently shaped season of the same size, and the response is to rephase rather than cut. Reforecasting the total while leaving the shape untouched silently assumes the first explanation every time.

When does reforecasting stop being useful?

When no remaining receipt can move. Once every delivery in the balance of the season is committed to production and cannot be rephased, reduced or cancelled, a reforecast produces a more accurate projection but not a different outcome. That is still worth doing for the financial view and for next season's learning, but it should be understood as reporting rather than as control, and it should not consume the same amount of meeting time.

Why record the reason a reforecast changed?

Because the variance narrative is the only durable output of the exercise. The revised numbers are superseded by the next reforecast within weeks; the record of what changed and why is what makes the next season's plan better. Without it, each season starts from last season's actuals with no memory of which figures were distorted by a late delivery, a stock-out or a one-off, and the same corrections get rediscovered annually.

How Often to Reforecast a Merchandise Plan

Reforecast cadence should follow the last point a decision can still change, not the reporting calendar. How to derive it, and what a reforecast must re-solve.

Cadence is a function of deadlines, not calendars

The question "how often should we reforecast" is usually answered with a reporting interval — monthly, because that is when the business closes. That is a reasonable default and it is not the right derivation.

The correct cadence is set by the last point at which a decision can still change. A reforecast has value only if something can be done differently as a result of it. If the next receipt is already committed to production, a reforecast produces a more accurate description of a season that is already settled.

Applying that test gives a range rather than a single answer:

Monthly is the floor for any business trading a seasonal calendar. Below that, a delivery can slip and be discovered two months later, by which point the escalation options have expired.
Weekly in season is right for direct-to-consumer channels, where sales data arrives continuously and receipts, reorders or cancellations remain available.
At each buy milestone, regardless of the regular rhythm. A reforecast immediately before a commitment can still change it; any number of reforecasts afterward can only describe it.
On any material event — a slipped delivery, a cut style, a channel change — rather than waiting for the next scheduled slot.

Reforecasting more often than decisions can be made generates meetings. Less often generates reports.

For the argument about what the in-season reforecast is really for — and why the number it produces matters less than the conversation it forces — see the perspective piece the in-season reforecast. This guide is the mechanics.

What a reforecast actually re-solves

A reforecast that updates actuals and stops looks finished, because every closed period now shows real numbers. That produces a plan that is accurate about the past and unchanged about the future — and the future is the half that decisions run on.

A reforecast in a merchandise financial plan has to run the same chain the original build ran, from the point of change forward:

Replace estimates with actuals for closed periods — sales, markdowns, receipts, and the resulting closing inventory.
Revise the phasing of the remaining periods. Not just the remaining total: the shape. This step is easy to skip, because a revised total still reconciles with the old shape left in place.
Re-derive inventory targets for each remaining period from the stock-to-sales ratio, which should also be checked rather than inherited — the remaining window is shorter than it was.
Re-solve receipts from the inventory identity for every remaining period, not only the one that changed.
Recalculate the open-to-buy position and reissue it, so buying is working against the current number rather than the pre-season one.
Record what changed and why.

Skipping step four is the specific error worth naming. Pushing a receipt from March to April and adjusting only April leaves the season total intact and every subsequent inventory target wrong. Each patch like that leaves the periods after it out of line with the inventory identity, and the errors stack until the plan no longer reconciles to itself — and it will keep balancing visually the whole time, because a spreadsheet balances whatever it is told to.

Smaller season, or differently shaped season?

This is the distinction that decides whether a reforecast helps or harms, and it is easy to skip because both situations present identically: sales are behind plan.

A smaller season means demand is genuinely lower than planned. The right response is to reduce receipts, cancel where possible, and plan the markdown exit earlier and shallower.

A differently shaped season means the same demand arriving on a different timeline — usually because deliveries landed late, a drop moved, or a marketing moment shifted. The right response is to rephase: move the remaining receipts to match the new shape, and hold the total.

These lead to opposite actions. Cutting receipts on a season that is merely late means being short when the demand arrives. Rephasing a season that is genuinely smaller means buying inventory that will clear at a discount.

A reforecast that revises the total while leaving the phasing untouched has silently chosen the first explanation, every time, without anyone deciding to. The corrective is to make the question explicit in the reforecast itself: is this less demand, or later demand, and what is the evidence? Full-price sell-through in the weeks the product was actually on the floor usually separates the two quickly — which is one of the practical reasons to phase full-price and markdown sales separately.

Keeping the narrative

The revised numbers a reforecast produces have a short life. The next reforecast supersedes them within weeks, and by season end nobody consults them.

The variance narrative — what changed, when, and why — is the part with a long life, and it exists only if someone writes it down, because the revised numbers carry no record of why they moved. Its value shows up a year later, when next season's plan is built from this season's actuals. Without the narrative, those actuals arrive as bare numbers: a soft week is just a soft week, indistinguishable from a week the product arrived late or the site was down or a competitor ran a promotion.

That is how a planning process comes to encode its own accidents. A week distorted by a stock-out becomes a permanent dip in the phasing curve. A late delivery becomes evidence that the category underperforms in that period. Each season inherits the previous season's noise as though it were signal, and the corrections get rediscovered rather than accumulated.

Recording a reason against each material change costs minutes per reforecast, and the return compounds, because every later plan built on those actuals inherits the explanation along with them. Because nothing downstream breaks without it, it is also the easiest thing to drop when the reforecast is running late.

When to stop

Reforecasting has a natural end point: when no remaining receipt can move.

Once every delivery in the balance of the season is committed to production and cannot be rephased, reduced or cancelled, the plan has no levers left. A reforecast at that point is worth producing for the financial view and for the record — the finance team still needs the projection, and next season still needs the narrative — but it should be recognised as reporting rather than control, and it should not occupy the same amount of anyone's week.

The useful discipline is to name that point in advance. Knowing the date after which the season is effectively fixed tells you how much of the calendar is genuinely decision-making time, and it tends to reveal that the window is shorter than the process assumes — which is an argument for reforecasting earlier and more often at the front of a season, and more lightly at the back.