BFCM Inventory and Fulfilment Ops for Shopify
The stock that runs out during BFCM is rarely the product you promoted. It is the free gift, the bundle component, and the variant nobody was watching.
On this page
Key takeaways
- Offer mechanics consume stock in patterns your forecast does not describe, because demand for a gift is driven by sales of something else.
- Continue-selling-when-out-of-stock is safe all year and dangerous during BFCM, when weeks of ordering compress into hours.
- Decide before the weekend what each offer does when its component runs out, because the alternative is deciding it at 2am.
- Fulfilment capacity is a hard constraint, and extending the sale window is the cheapest way to spread the load.
Most BFCM inventory planning forecasts demand for the products being promoted. That is the easy half, and it is not where stores get caught.
What runs out is the free gift attached to a threshold offer, the component shared between a bundle and its standalone listing, and the variant that was never the point of the campaign. These are consumed by demand for something else entirely, which is precisely why nobody is watching them.
Scope
Inventory and fulfilment operations for the BFCM period. Offer structure and checkout readiness are covered separately in the series.
Offer mechanics consume stock in unfamiliar patterns
Ordinary selling decrements stock in a way your forecast describes. Promotional mechanics do not, and each one fails differently.
Free gift with purchase. The gift is consumed at a rate driven by sales of the qualifying products, which may be your entire catalogue. If the threshold is easy to hit, the gift is effectively being ordered by every large basket.
Bundles. A component sold both inside a bundle and separately depletes from two directions. Fixed bundles compound this, because the bundle carries its own stock count that must stay synchronised with its components.
Buy X get Y. Drains the Y product at a rate governed by demand for X. If Y is a slower-moving item you chose precisely because it was slow, you may be about to discover how fast it can move.
Tiered thresholds. Shift the basket composition, so products that usually sell in ones start selling in twos and threes.
Mechanic and the stock it quietly consumes
| Mechanic | What actually depletes | Who is watching it |
|---|---|---|
| Free gift over threshold | The gift SKU | Usually nobody |
| Fixed bundle | Every component, plus bundle stock | Partly |
| Dynamic bundle | Components, from two directions | Rarely |
| Buy X get Y | The Y product | Rarely |
| Tiered spend | Higher quantities of mid-price items | No |
| Flat percentage | Broadly as forecast | Yes |
The rightmost column is the point. The mechanics that are hardest to forecast are also the ones nobody has been made responsible for watching, which is how a campaign runs out of its own centrepiece.
Decide the out-of-stock behaviour in advance
For every offer, there is a question with three possible answers, and you should choose deliberately rather than discover it.
When the gift, or the bundle component, runs out:
The offer stops applying. Cleanest, and the shopper sees the offer disappear. Requires that the mechanism actually checks stock, which not all do.
It substitutes. Another product takes its place. Good customer experience, needs configuring in advance, and needs a substitute chosen.
It oversells. The offer keeps applying and you owe people something you do not have. Occasionally acceptable if you can restock quickly; usually the worst outcome.
Set a stock alert on every gift, every bundle component and every Y product before the period starts. These are not the items your normal alerting covers, because they are not normally your constraint.
Inventory preparation
- Stock alerts on every gift, bundle component and buy-X-get-Y item
- A written decision for each offer about what happens at zero
- Continue-selling settings reviewed product by product, not globally
- Reserved allocation for anything central to an advertised campaign
- A check that bundle stock and component stock are actually synchronised
- Someone named as responsible for watching these during the period
Continue-selling is the setting that bites
Allowing sales when stock reaches zero is defensible for made-to-order products or where inbound stock is reliable and imminent. All year it causes no trouble.
BFCM changes the arithmetic without changing the setting. Volume that would normally arrive over weeks arrives in hours, so a product that would have oversold by three units across a fortnight oversells by three hundred over a weekend, and nobody notices until fulfilment.
Review this product by product before the period. It is tedious and it prevents the most expensive category of BFCM failure, which is taking money for things you cannot ship.
The damage is not only the cancellations. An oversold order that gets refunded a week later produces a support contact, a payment reversal, a customer who now distrusts you, and frequently a public review saying so — all attached to a customer you acquired at the highest cost of the year. The unit economics of an oversell during BFCM are far worse than the same mistake in March.
Fulfilment capacity is a real constraint
Inventory is only half of it. You also have to physically get parcels out, and your capacity to do that does not scale with demand.
Three things worth settling before the period:
Know your daily dispatch ceiling. Actual, not aspirational. How many parcels can be picked, packed and collected in a day with the people and space you have.
Compare it against your forecast. If peak-day orders exceed the ceiling, you will build a backlog, and a backlog during BFCM compounds because volume continues arriving.
Decide how to spread the load. Extending the sale window is the cheapest option — it moves demand off the peak days without reducing it. Early access for your email list does the same thing while rewarding your best customers. Both are decisions made in October, not during the weekend.
Communicate dispatch times honestly
A shopper told delivery will take a week and receiving it in a week is satisfied. One told two days and receiving it in a week is not, and contacts support about it. During BFCM the second scenario generates a support load on top of a fulfilment load, at the moment you have capacity for neither. Set the expectation wide.
Returns are part of the operation, not an afterthought
Return rates run higher on discounted and gifted purchases than on full-price ones, and the returns from BFCM arrive in December when everyone is exhausted.
Plan for the volume rather than being surprised by it. Make sure the returns process is documented and staffable, that returned stock can be received and re-shelved rather than accumulating in a corner, and that refunds are processed promptly — a slow refund after a discounted purchase is how a price-driven buyer becomes a permanent non-customer.
There is a specific complication with offer mechanics. When a customer returns part of an order that qualified for a gift or a threshold discount, the remaining order may no longer qualify. Decide your policy in advance — reclaim the gift, absorb it, or adjust the refund — and make sure whoever processes returns knows it. This question arrives during the first week of December without fail, and answering it ad hoc produces inconsistent decisions you will hear about later.
Forecasting when last year does not help
The standard method is to take last year's BFCM and adjust. It works when nothing changed, and something has almost always changed.
Three situations where last year is actively misleading:
You are running different mechanics. Last year's flat percentage tells you very little about how a threshold-plus-gift structure will consume stock. The total may be similar; the distribution across SKUs will not be.
Your catalogue has moved. Products added, discontinued, or repriced since last November change which items the offer pulls through. A bundle built around a product that was not in last year's range has no history at all.
Your traffic mix has changed. More paid, more email, more organic, a new market — each brings a different basket composition, and basket composition is what determines which components deplete.
What to do instead of ignoring history entirely: use last year for the aggregate, and model the distribution from the mechanic. For a gift over a threshold, estimate what share of orders will exceed that threshold based on your current average order value, and multiply. For a bundle, estimate the anchor product's uplift and derive component demand from it. These are rough numbers, and rough numbers that describe the right shape beat precise numbers describing the wrong one.
Buy the gift, not the forecast
For low-cost gift items, the cheapest insurance is to over-order deliberately. A surplus of a small promotional item costs you its unit price and some storage. Running out mid-campaign costs you the offer, the customer experience, and a decision made under pressure. The asymmetry is usually stark enough to make the answer obvious.
During the period
Watch component stock, not just headline products. The dashboard everyone looks at shows top sellers. The thing that will break is not on it.
Check the backlog daily. Orders placed against orders dispatched. A gap that grows two days running will not close on its own.
Cancel early if you must cancel. An oversell caught on day one and communicated clearly costs far less than a fulfilment failure discovered in the second week of December. Decide in advance who makes that call, because the instinct during peak is to hope it resolves itself.
For offer structure and what each mechanic costs, see the BFCM 2026 series. Inventory across retail and online is covered in POS and operations.
Frequently asked questions
- Why do my bundle components run out during BFCM?
- Because demand for a bundle component is driven by sales of the anchor product, not by its own standalone demand.
- Should I allow overselling during BFCM?
- Very rarely. Continue-selling-when-out-of-stock is defensible for made-to-order items or where you have reliable inbound stock.
- What happens when a free gift runs out mid-campaign?
- That depends entirely on your configuration, and most merchants have never checked. The possibilities are that the offer stops applying silently, that it substitutes something else, or that it oversells the gift.
- How far ahead should I plan BFCM inventory?
- Far enough that reordering is still possible, which for most supply chains means well before November.
- Should I extend my BFCM sale to spread fulfilment load?
- Often yes. A longer window spreads picking, packing and dispatch across more days, reduces the risk of a single-day failure destroying the period, and matches consumer behaviour, which now starts earlier.
Comments