Inventory Forecasting for Seasonal Wholesale: Planning Peaks Without Dead Stock

Seasonal demand is not difficult because it is unpredictable. It is difficult because the decision has to be made months before the evidence arrives: a buyer committing to Q4 volume in July is ordering against a number that will not be confirmed until November, and being wrong is asymmetric. Under-buying loses sales that were there; over-buying converts cash into cartons that sit until next year.

The answer is not a better prediction. It is a method that sizes the risk deliberately – a baseline from history, cover sized to real replenishment time, and buffers sized to how wrong the baseline usually is.

Start With a Baseline, Not a Guess

Every forecast starts from last season, adjusted. Three inputs build the baseline:

• Last season’s unit sales per SKU, cleaned. Remove weeks affected by stockouts, because a stockout week records demand as zero when it was not. This one correction fixes the most common seasonal forecasting error.
• The growth factor, taken from year-on-year sales outside the peak, since peak numbers already contain last year’s stockout distortion.
• The channel mix change. Adding a marketplace or region changes the shape of demand, not just its level, because channels peak on different dates.

Write the baseline as a weekly number, not a seasonal total: totals hide the shape, and the shape decides when stock must physically be in the warehouse. Keeping that weekly number honest across systems is covered in previsão de inventário e sincronização de estoque com um 3PL.

Lead Time Is the Multiplier That Matters

The most under-weighted number in seasonal planning is not demand, it is lead time – specifically full door-to-door lead time, not the production time the factory quotes:

• Production time, from confirmed order to finished goods ready.
• Material lead time, which in peak season often exceeds production time and is most often omitted.
• Inspection, freight, clearance and put-away, all of which stretch in Q4 before units are sellable.

Two consequences follow. The order date is set by working backwards from the week stock must be live, not forwards from today. And lead time variability deserves the same attention as demand variability, since a lead time that swings by two weeks needs the same buffer. Holiday season stock allocation covers how suppliers allocate capacity against those dates.

Sizing Safety Stock by Variability

Safety stock is not a comfort margin, it is a computed quantity: buffer scales with demand variability and with replenishment lead time. In plain terms:

• Stable demand, short lead time: a small buffer, often one to two weeks of cover.
• Stable demand, long lead time: a moderate buffer, because the exposure window is long even when demand is predictable.
• Variable demand, short lead time: a moderate buffer, because you can react quickly if the first read is wrong.
• Variable demand, long lead time: a large buffer – or, more realistically, a decision not to buy that SKU deep at all.

That last row is the real output of the exercise. Seasonal forecasting is often less about computing a number than about identifying SKUs whose risk cannot honestly be carried, and sourcing those through a faster lane instead. The trade-off between holding and financing that buffer is set out in ciclo de caixa de inventário de atacado, and the structural alternative – stocking only the predictable core – is the hybrid bulk inventory model.

The Seasonal Curve and the Pre-Season Commit

Peak demand is not a step change but a curve that rises, peaks and decays, and the shape differs by category, channel and geography:

• Commit the core, reserve the tail. Place a firm order for the confident portion – commonly sixty to seventy percent of the baseline – and reserve capacity for the rest, released against early-season actuals.
• Plan the peak week, not the season. Stock must be live two weeks before the peak week, or the peak itself is served from leftovers.
• Model the decay. Post-peak demand drops faster than intuition suggests, and late-season orders arrive into a falling market.

Reserved capacity is the part buyers neglect. Paying a small premium to hold an uncommitted slot turns one irreversible bet into two decisions, the second made with data. It pairs directly with verifying supplier capacity before peak season, because a reserved slot is only worth what the factory can deliver.

Dead Stock: Planning the Exit Before the Buy

Every seasonal buy should be placed with an exit already defined, because leftover stock is not inventory – it is cash with a storage bill:

• The clearance trigger. A date and remaining-quantity threshold at which markdown begins, set before the season so it is not negotiated emotionally later.
• The carry decision. Carrying stock is only rational when a year of storage and capital cost is below the markdown loss, which for bulky low-value goods it usually is not.

The operational reason this matters: space consumed by dead seasonal stock is unavailable for next season’s winners, a second-order cost that never appears in the markdown calculation. Buyers running several locations should also read distribuição multi-armazém, since splitting the buy across sites changes both the buffer and the clearance options. For SKUs where holding the buffer is simply uneconomic, prevenção de rupturas sem manter inventário sets out the alternatives.

The Review Rhythm

A seasonal forecast is a living document on a fixed calendar. Weekly during the season, compare actuals against baseline by SKU and act when cumulative variance exceeds the buffer – that is when a reserved-capacity release, an air top-up or an early markdown gets decided. Monthly outside the season, refresh lead times, which drift far more than demand does. The capacity and staffing constraints that bind once the forecast is set are covered in peak season readiness and cutoffs.

Contexto Técnico

Seasonal planning applies demand forecasting, where history is extrapolated and corrected for known distortions. The buffer protecting the forecast is a estoque de segurança quantity, a function of demand variability and replenishment lead time. The underlying decision – how much to buy when demand is uncertain and unsold units lose most of their value – is the classic newsvendor problem, which is why the exit plan belongs in the same document as the buy.

Perguntas Frequentes