Forecast demand without promotion and wholesale spikes
Keep dated sales unchanged, but label ordinary sales, promotions and exceptional wholesale orders. Build a recurring baseline from comparable trading periods, then add only events you expect to repeat, with their assumptions recorded. Discuss the resulting purchasing plan with Cambridge China Bridge.
Build a dated sales record
Export sales by product variant and channel, recording order date, dispatch date, units, selling price, customer reference and promotion reference. Choose a consistent date basis for analysis. Link returns and cancellations to the original sale so they do not appear as unrelated demand changes.
Add an event log showing promotion dates, unavailable stock, channel launches and exceptional wholesale orders. Preserve the original transactions and put classifications and adjustments in separate columns. Use units to plan purchases; keep revenue alongside them to explain price changes.
Classify the reason for each spike
Separate routine retail demand, recurring wholesale replenishment, promotional sales and exceptional wholesale orders. A large order is not automatically exceptional: ask whether it replenishes an established customer's stock or supplies a project that has ended. Mark unclear transactions for review rather than silently treating them as normal.
For promotions, record the offer, products, channels and selling dates. Keep genuine seasonal demand in the appropriate season. Look at sales before and after the event for purchases brought forward or shifted between variants; the event's sales total is not necessarily additional demand.
| Sales classification | Forecast treatment | Evidence to retain |
|---|---|---|
| Routine sales | Include in the recurring baseline | Comparable dated sales and stock availability |
| Recurring wholesale replenishment | Model separately if its timing differs from retail | Customer order history and expected replenishment dates |
| Promotion | Estimate ordinary demand during the event; add expected extra demand separately | Offer dates, comparable trading periods and future campaign plan |
| Exceptional wholesale order | Exclude from the recurring baseline; add a repeat only as a named event | Project purpose, customer confirmation and required delivery date |
Estimate ordinary demand during events
Compare promotion periods with non-promotional periods for the same variant and channel, allowing for season, trading days, price and stock availability. Estimate the ordinary demand that would have occurred during the event. Treat the difference from actual sales as an estimated event effect, with the comparison recorded.
Do not simply delete promotional periods: ordinary demand may still have occurred within them. Do not treat stockout sales as unrestricted demand either. Record any adjustment for unavailable stock separately, explain its basis, and leave uncertainty visible when comparable history is weak.
Write down what you expect to recur
For each forecast period, show the recurring baseline, planned promotional effect and named wholesale demand separately. Record the assumption, supporting sales or customer evidence, expected selling dates, person responsible and review date. For a promotion, explain whether the offer, channel and stock availability will resemble the historical event.
Keep an unconfirmed wholesale repeat in a separate scenario until there is supporting evidence. Show how purchases change if that event proceeds or disappears. When adding a confirmed customer order, remove any forecast allowance already covering it so the same demand is not counted again.
Working out your own order? Send us the product on WhatsApp and we reply the same working day: WhatsApp us
Turn the forecast into a purchasing brief
Translate expected selling dates into required stock arrival dates. Deduct usable stock and relevant incoming purchases, identify customer allocations, and show any stock buffer separately. Use reorder points for China stock for replenishment timing.
Give Cambridge China Bridge the product variants, dated demand breakdown, repeat assumptions and required arrival dates. Ask its staff in China to check the proposed purchasing schedule with factories. Use forecast versus firm order and supplier forecast and call-off terms when deciding what to communicate and commit. Compare actual sales with each demand category at the next review.
Frequently asked questions
Should I remove all promotional sales from my forecast?
No. Estimate the ordinary sales within the promotional period, then keep the extra promotional effect separate. Add that effect to future demand only where a planned event supports it.
How do I forecast a large wholesale order?
Check whether it is routine replenishment or a completed project. Keep exceptional demand outside the recurring baseline and record any expected repeat as a named event with supporting customer evidence.
What if I ran out of stock during a promotion?
Recorded sales show what you could supply. Flag the unavailable dates, use comparable available-stock periods where suitable, and record any estimate of missed demand separately.
How do I avoid counting customer orders twice?
Match confirmed orders to the product, customer and selling period they cover. Replace the corresponding forecast allowance rather than adding the whole order on top.